{"updated":"2026-10-15","listings":[{"ref":"MK-11664","title":"Multimodal human preference data (image, video, audio, 3D)","headline":"Pairwise human preference data for image, video, audio and 3D generation, from real people at global scale, plus on-demand collection","description":"Pairwise human preference and evaluation judgements for multimodal generative models, covering every modality except text (image, video, text-to-speech audio, 3D). Existing sets include ~300K text-to-video, ~2M text-to-image and ~300K text-to-speech preference pairs. Judgements come from real people (no synthetic raters) recruited through in-app placements via ad-network and mobile-developer partnerships, giving very high daily collection capacity worldwide for custom on-demand campaigns. Samples are available.","category":"Human preference / evaluation","sector":"AI & data","subsector":"Human preference / evaluation data collection","dataTypes":["Image","Video","Audio / speech","3D / CAD","Tabular / structured"],"useCases":["Preference / RLHF","Evaluation / benchmarks","Vision & perception","Speech & audio models"],"origin":"Purpose-collected (consented capture)","regions":["Global"],"languages":[],"sizeBand":"","count":"~2.6M existing preference pairs","volume":"Very high daily collection capacity; existing sets of ~300K video, ~2M image and ~300K text-to-speech preferences","years":"Ongoing","yearsOfHistory":"","growing":"Yes","systems":["Mobile ad / SDK placements"],"contents":["Text-to-video pairwise preferences (~300K)","Text-to-image pairwise preferences (~2M)","Text-to-speech pairwise preferences (~300K)","3D generation preferences (on demand)","Custom on-demand preference collection"],"formats":["Standard dataset formats"],"delivery":"","labelled":"Yes","labelTypes":["Pairwise human preferences"],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"Human preference / evaluation data collection","tools":[],"span":"","volumeHead":"2.6M existing preference pairs","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["human preference","rlhf","pairwise comparison","text-to-image","text-to-video","text-to-speech","multimodal","evaluation","on-demand collection"],"shortTitle":"Multimodal preference pairs"},{"ref":"MK-15049","title":"Large-scale microblog posts and follower graphs","headline":"~58B public microblog posts with full follower graphs, 2008-2026, JSON","description":"Crawled public microblog posts with complete follower graphs, 2008-2026. Fields include post text, likes, bookmarks, timestamp, author, author bio and follower relationships. The crawl focuses on influential accounts (mostly Western public figures and elites), so it is not a random platform-wide sample; English plus many other languages, US and worldwide. Seller pitches it for fine-tuning towards signal-dense writing.","category":"Text & documents","sector":"Marketing, media & publishing","subsector":"Social media archive","dataTypes":["Text","Tabular / structured"],"useCases":["Pre-training","Domain fine-tuning","Retrieval / RAG"],"origin":"Public / scraped sources","regions":["Global"],"languages":["English"],"sizeBand":"","count":"~58B posts","volume":"~58B posts","years":"2008–2026","yearsOfHistory":"18","growing":"","systems":["Web crawler","Cloud storage"],"contents":["Post text","Engagement counts (likes, bookmarks)","Timestamps","Author profiles and bios","Full follower graphs"],"formats":["JSON"],"delivery":"Cloud bucket (S3)","labelled":"Yes","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Public sources (per-source licences)","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"Social media archive","tools":[["files","File storage"]],"span":"18 years","volumeHead":"58B posts","sizeTB":null,"fromYear":2008,"toYear":2026,"yearsNum":18.0,"tags":["social media","microblog","follower graph","short-form text","crawl","pre-training","network data"],"shortTitle":"Microblog posts and follower graphs"},{"ref":"MK-21263","title":"Longitudinal company workforce panel, 1950-2020","headline":"Company-year workforce panel for 4M+ companies, 1950-2020, frozen for contamination-free forecasting benchmarks","description":"A longitudinal panel reconstructed from bulk-acquired professional career-history profiles, recording for each company and year the observed workforce, role structure (10 role buckets), capability buckets and granular skills, with quality tiers. The panel is frozen at 2020 so post-2020 outcomes are verifiable but absent from the inputs, supporting contamination-controlled forecasting benchmarks, backtests and training environments. It ships with derived event labels (workforce contraction, growth reversal) with known outcomes, curated reference cases and pre-registered evaluation cohorts with frozen predictions. An optional pseudonymised person-level layer (employer, title, normalised role, start/end month, skills, coarse location) underlies the aggregates; the seller notes survivorship/hindsight bias, observed-profile counts rather than payroll headcount, and thinner coverage before 2000 and for blue-collar roles. A reference-forecasting engine and evaluation protocol can also be supplied.","category":"Workforce & labour market","sector":"Staffing & HR","subsector":"Labour-market / company workforce panel","dataTypes":["Tabular / structured"],"useCases":["Forecasting & time series","Evaluation / benchmarks","RL environments","Domain fine-tuning","Retrieval / RAG"],"origin":"Licensed from third parties","regions":["Global"],"languages":["English"],"sizeBand":"","count":"~48M company-years (4M+ companies)","volume":"4M+ companies; 48M company-years; 1.88B skill rows; 46.5M capability buckets","years":"1950–2020","yearsOfHistory":"70","growing":"No","systems":["Database"],"contents":["Company-year records (company ID, industry, country, year, observed profiles, growth)","Role shares across 10 role buckets","Capability buckets (~46.5M)","Granular skill rows (~1.9B)","Quality tier per company-year","Entity-type flags (placeholder and non-business pages)","Derived event labels with known outcomes (workforce contraction, growth reversal)","Curated reference cases","Pre-registered evaluation cohorts with frozen predictions","Optional pseudonymised person-level career records"],"formats":["Parquet","DuckDB","CSV"],"delivery":"SFTP; Cloud bucket; API / MCP query access (evaluation)","labelled":"Yes","labelTypes":["Quality tiers","Entity-type flags","Event labels with known outcomes","Reference cases","Pre-registered evaluation cohorts"],"personalData":"Moderate","deidentification":"Done by owner","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"Labour-market / company workforce panel","tools":[["database","Database"]],"span":"70 years","volumeHead":"48M company-years","sizeTB":null,"fromYear":1950,"toYear":2020,"yearsNum":70.0,"tags":["workforce","labour market","company panel","skills","career histories","forecasting","benchmark","backtesting","longitudinal","parquet"],"shortTitle":"Company workforce panel since 1950"},{"ref":"MK-26146","title":"Financial-research software company: 5-year archive","headline":"Financial-research software company: 5 years of email, chat, 100+ repos, issues, wiki, CRM and ~200M-row warehouse","description":"Internal data of a financial-data and research software company serving banks and asset managers, ~0.7 TB of known size over about five years. Includes business email and Drive files, ~270K team-chat messages across ~150 public channels, a community server with ~70 channels, 100+ code repositories with ~3K issues, issue-tracker work items, internal wiki documentation, ~160K CRM contacts, application database records, a data warehouse of ~370 tables and ~200M rows (product analytics, user DB, sandbox), object storage and product design files.","category":"Company operations archive","sector":"Software & internet","subsector":"Financial research and analytics software","dataTypes":["Email","Documents","Chat & messaging","Code","Tabular / structured"],"useCases":["Coding models","Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"100 GB-1 TB","count":"~200M warehouse rows","volume":"~0.7 TB of known size; ~270K chat messages; ~160K CRM contacts; 100+ repos / ~3K issues; ~370 warehouse tables / ~200M rows","years":"2021–2026","yearsOfHistory":"5","growing":"Yes","systems":["Email","File storage","Team chat","Community chat","Git","Issue tracker","Wiki","CRM","Database","Data warehouse","Cloud storage","Design files"],"contents":["Business email and Drive files","Team chat messages (~270K)","Community server channels","Code repositories (100+) and issues (~3K)","Issue-tracker work items","Internal wiki documentation","CRM contacts (~160K)","Application database records","Data warehouse (~370 tables, ~200M rows)","Product design files"],"formats":["Chat, issue-tracker and wiki extracts","Mixed"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Financial research and analytics software","tools":[["chat","Team chat"],["tickets","Tickets"],["git","Git"],["email","Email"],["crm","CRM"],["wiki","Docs & wiki"],["database","Database"],["design","Design files"],["files","File storage"]],"span":"5 years","volumeHead":"200M warehouse rows","sizeTB":0.67,"fromYear":2021,"toYear":2026,"yearsNum":5.0,"tags":["fintech","financial software","codebase","team chat","crm","data warehouse","wiki","issue tracker"],"shortTitle":"Financial-research software archive"},{"ref":"MK-26982","title":"TV and film viewing data, metrics and title embeddings","headline":"TV and film viewing data: household viewing history, title metrics, ~84K title embeddings and scored expert workflows","description":"Four linked assets from a TV and film audience-measurement business: (1) household-level viewing history across 230+ countries and territories (household ID, timestamp, title/episode, country, locality, connection class) in daily compressed files, ~0.9 TB, 2017-2026; (2) title and platform metrics from a ~2.3 TB analytics database (reach, demand indices, demand spikes, local and audience fit, library lift, season retention, talent demand) with definitions, units and dates; (3) trained title embeddings for ~84K titles; and (4) expert business workflows for content licensing, platform-growth campaigns, media planning and streaming-company research, with questions, evidence, worked answers, grading rules and an automatic scorer, plus tool (MCP) access. Shared title IDs join all four.","category":"Consumer behaviour","sector":"Marketing, media & publishing","subsector":"TV and film audience measurement","dataTypes":["Tabular / structured","Text"],"useCases":["Forecasting & time series","RL environments","Evaluation / benchmarks","SFT / instruction tuning","Computer-use & tool-use agents","Domain fine-tuning"],"origin":"First-party company records","regions":["Global"],"languages":[],"sizeBand":"1-10 TB","count":"~84K titles","volume":"~3.2 TB (viewing ~0.9 TB; analytics database ~2.3 TB; embeddings 175 MB)","years":"2017–2026","yearsOfHistory":"9.5","growing":"Yes","systems":["Database","Daily data files","Model files","Methodology documents","MCP server"],"contents":["Household viewing histories (daily files)","Title and platform metrics with definitions","Country and genre demand data","Title embeddings (~84K titles)","Expert workflow tasks with worked answers","Grading rules and automatic scorer","Sample training tasks"],"formats":["CSV.gz","CSV","JSON","NumPy","JSONL"],"delivery":"Download; API","labelled":"Partly","labelTypes":["Metric definitions and units","Workflow scoring rules"],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"TV and film audience measurement","tools":[["database","Database"]],"span":"9.5 years","volumeHead":"84K titles","sizeTB":3.2,"fromYear":2017,"toYear":2026,"yearsNum":9.5,"tags":["tv","film","streaming","viewing history","audience measurement","embeddings","recommendations","rl tasks","media"],"shortTitle":"TV and film viewing data"},{"ref":"MK-30328","title":"Food-service distribution centre: ERP, email and files","headline":"Closed food-service distribution centre: 13 years of two ERPs, email and SOP/ISO, customs and contract files","description":"Records of a former national food-service distribution centre for a global quick-service restaurant brand (imports, receiving, quality checks and delivery to restaurants), which ceased operations in 2024. Includes two generations of ERP (2013-2018 and 2018-2024) covering general ledger, bank transactions and reconciliations, monthly income statements, customer invoicing and deliveries to restaurant outlets, purchase orders, goods receipts and supplier payments, and item master, stock movements and valuation; 13 years of email with the brand, outlets and suppliers (orders, customs and import paperwork, quality and delivery issues, contracts) in Arabic and English; and shared folders with contracts, SOPs, quality-control checklists, ISO and audit material, and customs and shipping documents.","category":"Company operations archive","sector":"Logistics, distribution & wholesale","subsector":"Food-service distribution centre","dataTypes":["Tabular / structured","Email","Documents"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Forecasting & time series","Retrieval / RAG"],"origin":"First-party company records","regions":["Middle East"],"languages":["Arabic","English"],"sizeBand":"10-100 GB","count":"","volume":"51 GB (email 30 GB; ERP 18.7 GB; files 2 GB)","years":"2011–2024","yearsOfHistory":"13","growing":"No","systems":["ERP","Email","File storage"],"contents":["General ledger and bank reconciliations","Monthly income statements","Sales invoices and deliveries to restaurant outlets","Purchase orders, goods receipts and supplier payments","Item master, stock movements and valuation","Email with brand, outlets and suppliers","Customs declarations and shipping documents","SOPs and quality-control checklists","ISO documentation and audit material","Supplier and customer contracts"],"formats":["SQL extract","CSV","PST","File copy"],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Food-service distribution centre","tools":[["email","Email"],["erp","ERP"],["files","File storage"]],"span":"13 years","volumeHead":"51 GB","sizeTB":0.051,"fromYear":2011,"toYear":2024,"yearsNum":13.0,"tags":["food distribution","supply chain","erp","quality control","customs","sop","email archive","arabic"],"shortTitle":"Food-service distributor records"},{"ref":"MK-37223","title":"DTC health supplements brand: 200K+ orders and full growth playbook","headline":"DTC supplements brand: 200K+ orders in 14 months with marketing, ad-platform, contract and operations data","description":"Operating records of a US direct-to-consumer health and wellness supplements brand that sold 200,000+ orders in 14 months. Includes e-commerce orders, email and paid-acquisition marketing data across major social, search and app ad networks, signed contracts and partner documents, payroll/banking records, and files covering who the brand worked with and how operations were run (~15 TB stated, about two years). Files live in a cloud drive, servers and team chat.","category":"Consumer behaviour","sector":"Retail & e-commerce","subsector":"DTC health supplements brand","dataTypes":["Tabular / structured","Documents","Chat & messaging","Image","Video"],"useCases":["Enterprise workflow agents","Forecasting & time series","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"10-100 TB","count":"200K+ orders","volume":"~15 TB; 200,000+ orders","years":"2026","yearsOfHistory":"2","growing":"Yes","systems":["E-commerce","Email marketing","Ad platforms","E-signature","HR & payroll","File storage","Team chat"],"contents":["E-commerce orders (200K+)","Email marketing campaigns and flows","Paid ads data across several ad networks","Contracts and partner documents","Payroll and banking records","Growth and operations playbook documents"],"formats":["Platform exports","Documents"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"DTC health supplements brand","tools":[["chat","Team chat"],["shop","Commerce"],["hr","HR & payroll"],["analytics","Analytics & marketing"],["sign","E-signature"],["files","File storage"]],"span":"2 years","volumeHead":"200K orders","sizeTB":15.0,"fromYear":null,"toYear":2026,"yearsNum":2.0,"tags":["dtc","supplements","ecommerce","orders","paid ads","email marketing","growth","contracts"],"shortTitle":"DTC supplements brand playbook"},{"ref":"MK-37963","title":"Humanoid and bimanual robot teleoperation data","headline":"Robot teleoperation data: ~700 h humanoid, ~50 h bimanual arms, ~500 h handheld UMI, scaling to ~5,000 h/month","description":"Teleoperation data collected in Malaysia from a humanoid robot (~600 hours upper-body manipulation and ~100 hours whole-body loco-manipulation), bimanual arm stations (~50 hours, ~20 stations just started) and handheld universal-manipulation-interface (UMI) rigs (~500 hours, whole-body). Existing hours are available off the shelf and collection is scaling towards ~5,000 hours per month. Samples are available.","category":"Robotics","sector":"AI & data","subsector":"Robotics teleoperation data collection","dataTypes":["Video","Sensor / telemetry"],"useCases":["Robotics / embodied AI","World models"],"origin":"Purpose-collected (consented capture)","regions":["Southeast Asia"],"languages":[],"sizeBand":"","count":"~1,250 hours","volume":"~700 hours humanoid (600 upper-body, 100 whole-body); ~50 hours bimanual; ~500 hours UMI; scaling to ~5,000 hours/month","years":"Ongoing","yearsOfHistory":"","growing":"Yes","systems":["Teleoperation rigs","UMI rigs"],"contents":["Humanoid upper-body manipulation teleop","Humanoid whole-body loco-manipulation teleop","Bimanual arm-station teleop","Whole-body UMI recordings"],"formats":["Teleop / UMI recordings"],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"Robotics teleoperation data collection","tools":[],"span":"","volumeHead":"1.2K hours","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["robotics","teleoperation","humanoid","bimanual","umi","manipulation","loco-manipulation","embodied ai"],"shortTitle":"Humanoid teleoperation data"},{"ref":"MK-44927","title":"Indian food-processing machinery maker: job folders since 2007","headline":"Indian assembly-line machinery maker: 5K+ job folders with quotes, CAD drawings, photos and signed work orders","description":"Records of an Indian manufacturer of assembly-line machines for food-processing companies. The core is 5K+ job folders going back to 2007, each holding a quote, CAD drawings, photos and a signed work order, plus years of inspection reports and PDFs. The company also holds about 5 years of CRM customer data and about 10 years of Drive files covering designs, payments and the general ledger. About 4 TB in total, stored on a file server.","category":"Company operations archive","sector":"Manufacturing","subsector":"Assembly-line machinery for food processing","dataTypes":["Documents","CAD / 3D","Image","Tabular / structured"],"useCases":["Domain fine-tuning","Vision & perception","Retrieval / RAG","Enterprise workflow agents"],"origin":"First-party company records","regions":["South Asia"],"languages":[],"sizeBand":"1-10 TB","count":"5K+ job folders","volume":"~4 TB; 5K+ job folders (quote, CAD drawings, photos, signed work order each); inspection reports and PDFs; ~5 years of CRM customer data; ~10 years of Drive files","years":"2007","yearsOfHistory":"","growing":"","systems":["File storage (Google Drive)","CRM (HubSpot)","Accounting"],"contents":["Job folders with quotes, CAD drawings, photos and signed work orders","Inspection reports","Designs and project files","Payment records and general ledger","CRM customer data"],"formats":["CAD","Photos","PDF"],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"Mid-size company","employees":"51–200","industry":"Assembly-line machinery","tools":[["hubspot","HubSpot"],["googledrive","Google Drive"],["accounting","Accounting"]],"span":"10 years","volumeHead":"5K job folders","sizeTB":4.0,"fromYear":2007,"toYear":null,"yearsNum":null,"tags":["manufacturing","food processing","machinery","cad","quotes","work orders","inspection reports","india"],"shortTitle":"Food-machinery job folders and CAD"},{"ref":"MK-51803","title":"Physical-contact annotations for robot assembly recordings","headline":"Contact annotations (contact/no-contact, confidence, uncertainty, force-sensor references) keyed to real robot assembly recordings: ~41.7K rows on 40 recordings","description":"Proprietary physical-contact labels keyed to real robot assembly recordings: contact/no-contact, confidence, uncertainty and force-sensor references. First release covers 40 recordings of a public research dataset (CC BY 4.0) with ~41.7K annotation rows.","category":"Robotics","sector":"AI & data","subsector":"Robotics annotation","dataTypes":["Sensor / telemetry","Video"],"useCases":["Robotics / embodied AI"],"origin":"Expert-created","regions":["Europe"],"languages":[],"sizeBand":"","count":"~41.7K rows","volume":"40 recordings; ~41.7K annotation rows","years":"2025 source; 2026 labels","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":["Parquet","JSONL"],"delivery":"","labelled":"Yes","labelTypes":[],"personalData":"None","deidentification":"Not needed","rights":"Owner-held","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"Robotics annotation","tools":[],"span":"","volumeHead":"42K rows","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["robotics","contact","force","assembly","annotation"],"shortTitle":"Robot assembly contact labels"},{"ref":"MK-54760","title":"Legal e-discovery evaluation matters (attorney-built, synthetic)","headline":"Attorney-built legal matter for e-discovery agents: 1,486 emails + 38 exhibits with planted evidence chains; new matters built to order","description":"An attorney-built, fully fictional legal matter for evaluating legal agents: a custodian email corpus and exhibit set with planted evidence chains across ten arcs, testing responsiveness, privilege, chronology, knowledge analysis and multi-document joins. Recent runs show frontier models scoring between ~4% and ~44%. The owner builds new matters to a buyer's specification in 3-7 days (reviewed benchmark in 1-3 weeks) with a team of attorneys.","category":"RL environment / benchmark","sector":"Legal","subsector":"Legal AI evaluation","dataTypes":["Email","Documents","Tabular / database"],"useCases":["Evaluation / benchmarks","Legal agents","Long-context reasoning"],"origin":"Expert-built (synthetic)","regions":["North America"],"languages":["English"],"sizeBand":"<10 GB","count":"1,486 emails + 38 exhibits","volume":"1,486 emails + 38 exhibits (~31 MB) per matter; new matters to order","years":"Not applicable","yearsOfHistory":"","growing":"Yes","systems":[],"contents":["Fictional matter's email corpus with custodian metadata","38 PDF exhibits (agreements, easements, filings, audit materials)","planted evidence chains (10 arcs) for responsiveness, privilege, chronology, knowledge analysis and multi-document joins","graded task set"],"formats":["EML","PDF","CSV","DAT","PNG"],"delivery":"Download","labelled":"Yes","labelTypes":["Responsiveness","privilege","evidence chains","chronology"],"personalData":"None","deidentification":"Not applicable (fictional names)","rights":"Owner-held","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"Legal AI evaluation","tools":[],"span":"","volumeHead":"1.5K emails + 38 exhibits","sizeTB":3e-05,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["legal","e-discovery","privilege review","evaluation","synthetic","seed world"],"shortTitle":"Legal e-discovery evals"},{"ref":"MK-58833","title":"Household AI-assistant startup: 2 years of ops sheets, Slack, Notion, GitHub and a 200 GB+ database with AI eval sets","headline":"Indian AI kitchen-management and remote-butler service: operations, accounting and estimation sheets, Slack, Notion product docs, codebase, ~7,000 Drive files and a 200 GB+ database including AI evaluation sets","description":"Operating record of an AI household-assistant startup: operations, accounting and estimation spreadsheets, team chat, product documentation, codebase, ~7,000 shared files and a production database that includes in-house AI evaluation sets.","category":"Company operations archive","sector":"Consumer services","subsector":"Home services / AI assistant","dataTypes":["Code","Team chat","Documents","Tabular / structured","Evaluation sets"],"useCases":["Enterprise workflow agents","Agent evaluation"],"origin":"First-party company records","regions":["South Asia"],"languages":[],"sizeBand":"100 GB-1 TB","count":"","volume":"~500 GB; 35 mailboxes; ~7,000 Drive files","years":"~2 years","yearsOfHistory":"","growing":"","systems":["Google Sheets","Slack","Notion","GitHub"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Medium","deidentification":"","rights":"Owner-held","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Home services / AI assistant","tools":[["slack","Slack"],["github","GitHub"],["notion","Notion"],["googlesheets","Google Sheets"]],"span":"2 years","volumeHead":"500 GB","sizeTB":0.5,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["startup archive","ai assistant","eval sets","operations"],"shortTitle":"Household AI-assistant startup"},{"ref":"MK-59134","title":"Multi-site dog daycare, boarding and grooming chain: 8-year customer-journey archive","headline":"Multi-site dog daycare and boarding chain: 8 years linking calls, texts, tickets and POS visits for the same customers","description":"Operating data of a US dog daycare, boarding and grooming business with 10+ locations and a central call centre, ~0.4 TB already exported. The customer journey is linked end to end, from lead intake to booked meet-and-greet to enrolment and ongoing service: CRM with 93.6K contacts, 47.6K deals, 2.6M service tickets, 2.0M tasks, 503K call logs with 66K call recordings, 327K emails, 411K texts, 596K inbox messages and 436K notes, plus a phone system with 33.9K calls and 26.2K recordings. Point-of-sale records cover 27K customer accounts, 31K pets, 540K daycare/boarding visits, 654K orders, vaccinations and service contracts (35.7M rows). Also 11.2K customer agreements and waivers (7.9K signed) with field-level data, 168K timesheets, 176K shifts and 1,464 employee records, ~193K team-chat messages, 119K Drive files (SOPs, training, store P&Ls, financial models), and an in-house analytics warehouse (57 tables, 11.5M rows incl. 2.96M purchase records) with its dashboard code and logs.","category":"Company operations archive","sector":"Consumer services","subsector":"Dog daycare, boarding and grooming chain","dataTypes":["Audio / speech","Text","Chat & messaging","Email","Tabular / structured","Documents","Code"],"useCases":["Enterprise workflow agents","Speech & audio models","Forecasting & time series","Domain fine-tuning","Computer-use & tool-use agents"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"100 GB-1 TB","count":"2.6M service tickets","volume":"~0.4 TB exported; 93.6K CRM contacts; 47.6K deals; 2.6M service tickets; 66K recorded calls (CRM) + 26K (phone system); 411K texts; 596K inbox messages; 540K visits; 654K orders; 35.7M POS rows; ~193K chat messages; 119K documents; 168K timesheets","years":"2018–2026","yearsOfHistory":"8","growing":"Yes","systems":["CRM","POS","Telephony","Email","File storage","Team chat","E-signature","Scheduling / timekeeping","Payroll","Data warehouse","Accounting","Expense management","Ads"],"contents":["CRM contacts, deals and service tickets (2.6M)","Recorded customer calls (66K CRM + 26K phone system)","Text messages (411K) and inbox messages (596K)","Meet-and-greet bookings and enrolments","Daycare/boarding visits (540K) and orders (654K)","Pet records and vaccinations","Signed customer agreements and waivers","Timesheets, shifts and employee records","Store P&Ls and financial models","SOPs and training material","Team chat (~193K messages)","Analytics warehouse tables, dashboard code and logs"],"formats":["CSV","JSONL","SQLite","Parquet","MP3","MBOX","Office files","Signed PDFs","Code repo mirror"],"delivery":"Download","labelled":"Partly","labelTypes":["AI call summaries","Call tags and comments"],"personalData":"High","deidentification":"","rights":"Owner-held","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"Large organisation","employees":"200+","industry":"Dog daycare, boarding and grooming chain","tools":[["chat","Team chat"],["email","Email"],["crm","CRM"],["calls","Calls"],["database","Database"],["accounting","Accounting"],["shop","Commerce"],["hr","HR & payroll"],["analytics","Analytics & marketing"],["calendar","Calendar"],["sign","E-signature"],["files","File storage"]],"span":"8 years","volumeHead":"2.6M service tickets","sizeTB":0.4,"fromYear":2018,"toYear":2026,"yearsNum":8.0,"tags":["pet care","dog daycare","call recordings","customer journey","crm","pos","call center","multi-location"],"shortTitle":"Dog daycare chain archive"},{"ref":"MK-62641","title":"US e-commerce product and price data from 3,500 retailers, mapped to major marketplaces","headline":"US e-commerce product and price data from ~3,500 retailers, cross-mapped to the largest marketplaces (10-15+ TB)","description":"Product and price data collected from ~3,500 US retail websites via retailers, crawlers and APIs, aggregated and cross-checked against the three largest US marketplaces, with mappings between sources (retailer, marketplace product identifiers and listing metadata). Also includes price-drop history, consumer logs, API and MCP call data and raw data. A schema manifest can be shared; total volume 10-15+ TB from 2026.","category":"Consumer behaviour","sector":"Retail & e-commerce","subsector":"E-commerce product and price aggregation","dataTypes":["Tabular / structured","Text"],"useCases":["Retrieval / RAG","Computer-use & tool-use agents","Forecasting & time series","Pre-training","Domain fine-tuning"],"origin":"Public / scraped sources","regions":["North America"],"languages":[],"sizeBand":"10-100 TB","count":"~3,500 retailer sites","volume":"10-15+ TB","years":"2026","yearsOfHistory":"","growing":"","systems":["Web crawlers","APIs","Database"],"contents":["Product records from ~3,500 retailers","Prices and price-drop history","Cross-marketplace product mappings and identifiers","Marketplace listing metadata","Consumer logs","API and MCP call logs","Raw collected data"],"formats":["CSV","JSON"],"delivery":"API; Cloud bucket (S3); FTP","labelled":"Partly","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"E-commerce product and price aggregation","tools":[["database","Database"]],"span":"","volumeHead":"3.5K retailer sites","sizeTB":15.0,"fromYear":2026,"toYear":2026,"yearsNum":null,"tags":["ecommerce","pricing","product catalog","price history","marketplaces","product matching","shopping agents","usa"],"shortTitle":"US retail product and price data"},{"ref":"MK-63745","title":"Multi-location bakery chain: POS, operations warehouse, email and team chat (7 years)","headline":"Multi-location bakery: ~1.06M line-item POS orders, an ops warehouse with a demand-prediction model, ~150K emails and ~54K team messages","description":"Seven years of a multi-location bakery with central production, catering and wholesale (~$2M revenue): line-item POS orders since 2020, a Postgres + dbt operations warehouse since 2022 (production lists, end-of-night counts, waste, inventory, till audits and a bake-prediction model), wholesale and catering orders, the customer-service mailbox foldered by request type, the owner's mailbox, team chat since 2019, ops meeting notes with transcripts, SOPs and a log of filed decisions, documents and accounting.","category":"Company operations archive","sector":"Hospitality & food","subsector":"Bakery and food retail chain","dataTypes":["Tabular / database","Email","Team chat","Documents","Code"],"useCases":["Enterprise / business agents","Forecasting","Customer-service agents","SFT / instruction tuning"],"origin":"First-party company records","regions":["North America"],"languages":["English"],"sizeBand":"10-100 GB","count":"~1.06M POS orders","volume":"~1.06M POS orders; ~36K + ~115K emails; ~54K chat messages; ~2,400 documents; ~80 dbt models; ~35-50 GB total","years":"2019–2026","yearsOfHistory":"7","growing":"Yes","systems":["POS","Postgres + dbt warehouse","Microsoft 365 email","Team chat","Notion","Google Drive","Accounting"],"contents":["~1.06M POS orders at line-item level (2020-)","ops warehouse (bake and dough lists, end-of-night counts, waste, inventory, till audits, bake-prediction model","~80 dbt models, ~1,100 commits)","wholesale and catering orders + ~1,400 catering emails","customer mailbox ~36K messages (~19K replies, foldered by type)","owner mailbox ~115K","~54K team-chat messages","ops meeting notes, SOPs and ~100 filed decisions","~2,400 Drive files","accounting 2020-"],"formats":["Database exports","SQL/dbt","Email","JSON","Documents"],"delivery":"Download","labelled":"Partly","labelTypes":["Email folders by type","filed decisions"],"personalData":"Moderate","deidentification":"Owner will de-identify via a third party","rights":"Owner-held","sample":"On request","readiness":"Ready now","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Bakery and food retail chain","tools":[["email","Outlook"],["notion","Notion"],["googledrive","Google Drive"],["postgresql","Postgres"],["chat","Team chat"],["accounting","Accounting"],["shop","Commerce"]],"span":"7 years","volumeHead":"1.1M POS orders","sizeTB":0.05,"fromYear":2019,"toYear":2026,"yearsNum":7.0,"tags":["food retail","POS","operations","demand forecasting","customer service","SOPs","dbt"],"shortTitle":"Bakery chain operations"},{"ref":"MK-66152","title":"Point-of-care diagnostics R&D archive","headline":"Point-of-care diagnostics R&D archive: experiments, raw data, videos, CAD, 16 code repos, email and R&D chat","description":"R&D record (2020-2026) of a small point-of-care diagnostics company working on plasma separation, biomarker detection, experimental analysis and assay simulation. A shared drive (~52 GB, 411 items) holds 70 videos, ~216 images/vector files, reports, CAD, spreadsheets and research records; 16 private code repositories (~36.5K lines, 444 commits, 30 PRs) cover assay simulation, analysis, microscopy assets and web/mobile apps; curated protocols, raw XLSX/CSV instrument data, Python analysis, figures and design-history reports; an experiment tracker with 71 recorded test entries across several experiments; plus ~2K emails and an estimated 5-10 GB of R&D chat. Includes some third-party reference material.","category":"Healthcare & life sciences","sector":"Healthcare & life sciences","subsector":"Point-of-care diagnostics R&D","dataTypes":["Documents","Code","Video","Image","3D / CAD","Tabular / structured","Email","Chat & messaging"],"useCases":["Domain fine-tuning","Coding models","Retrieval / RAG","Evaluation / benchmarks"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"10-100 GB","count":"~411 drive items","volume":"~53 GB plus est. 5-10 GB chat text; 411 drive items; 16 repos / 444 commits; 71 recorded test entries","years":"2020–2026","yearsOfHistory":"6","growing":"","systems":["File storage","Git","Spreadsheets","Email","Team chat"],"contents":["Experiment protocols and plans","Raw XLSX/CSV experimental data","Experiment tracker (71 test entries)","Lab videos (70) and microscopy images","CAD files","Private code repositories (16)","Python analysis and figures","Design-history reports","R&D email (~2K)","R&D chat messages"],"formats":["Drive files","Git","XLSX","CSV","Video","CAD"],"delivery":"","labelled":"Partly","labelTypes":["Experimental records with replicates, controls and reference measurements"],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"Point-of-care diagnostics R&D","tools":[["chat","Team chat"],["git","Git"],["email","Email"],["sheet","Spreadsheets"],["files","File storage"]],"span":"6 years","volumeHead":"411 drive items","sizeTB":0.06,"fromYear":2020,"toYear":2026,"yearsNum":6.0,"tags":["biotech","diagnostics","point of care","lab experiments","r&d","assay","cad","code repositories"],"shortTitle":"Diagnostics R&D archive"},{"ref":"MK-67721","title":"Bowling and events venue: bookings and communications","headline":"Long-established bowling and events venue: ~5,200 reservations (2023-2026), party enquiries and customer email threads","description":"A long-established bowling venue's operating records: ~5,200 deduplicated reservations in its booking system from Dec 2023 to Oct 2026 (dates, times, party sizes), including 230+ birthday-party bookings, party and event enquiries and the booking process by email, customer conversations, party-planning and staff-scheduling messages and notes to the team. An anonymised sample of email threads (HTML, TXT, JSON) and booking-system screenshots is available. Payroll is excluded; security-camera footage may also be offered.","category":"Company operations archive","sector":"Consumer services","subsector":"Bowling and party/events venue","dataTypes":["Tabular / structured","Email","Chat & messaging","Documents"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Forecasting & time series"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"<10 GB","count":"~5,200 reservations","volume":"~5 GB; ~5,200 reservations (Dec 2023-Oct 2026), 230+ birthday-party bookings","years":"2023–2026","yearsOfHistory":"3","growing":"Yes","systems":["Email","File storage","Payroll","Team chat"],"contents":["Reservation records (date, time, party size)","Birthday-party and event enquiries","Customer email conversations","Weekly party-planning messages","Staff scheduling","Notes to the team","Payroll records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"Bowling and party/events venue","tools":[["chat","Team chat"],["email","Email"],["hr","HR & payroll"],["files","File storage"]],"span":"3 years","volumeHead":"5.2K reservations","sizeTB":0.005,"fromYear":2023,"toYear":2026,"yearsNum":3.0,"tags":["bowling","events","reservations","party bookings","small business","scheduling","hospitality"],"shortTitle":"Bowling and events venue"},{"ref":"MK-69567","title":"Global street-level feature map: 780M geolocated, labelled objects","headline":"Global street-level feature map: 780M geolocated, finely labelled real-world objects from imagery and video (2014-2026)","description":"About 780M real-world features detected in a global index of street-level imagery and video: infrastructure, branded signage and vehicles, vegetation and road conditions. Each record carries a fine-grained label (~10M label classes), latitude/longitude, address and capture date, machine-labelled with a fine-tuned vision model. Sources are open mapping sites plus the owner's own capture, covering 184 countries from 2014 to 2026. Suited to long-tail vision and multimodal training, geolocation and world models, robotics perception and change detection, and location-grounded retrieval. About 99% of the source imagery comes from an open street-level platform under CC BY-SA (share-alike), about 1% from the owner's own capture; 96%+ stated label accuracy and ~220K 360-degree videos.","category":"Geospatial & imagery","sector":"AI & data","subsector":"Street-level geospatial feature mapping","dataTypes":["Geospatial","Tabular / structured"],"useCases":["Vision & perception","World models","Robotics / embodied AI","Retrieval / RAG","Pre-training"],"origin":"Public / scraped sources","regions":["Global"],"languages":[],"sizeBand":"","count":"780M feature records","volume":"780M rows","years":"2014–2026","yearsOfHistory":"12","growing":"","systems":[],"contents":["Geolocated feature records (780M)","Fine-grained labels (~10M classes)","Latitude and longitude","Street address","Capture date","Infrastructure, signage, vehicle, vegetation and road-condition features"],"formats":["CSV"],"delivery":"Cloud bucket (S3)","labelled":"Yes","labelTypes":["Machine-generated fine-grained object labels"],"personalData":"Low","deidentification":"","rights":"Public sources (per-source licences)","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"Street-level geospatial feature mapping","tools":[],"span":"12 years","volumeHead":"780M feature records","sizeTB":null,"fromYear":2014,"toYear":2026,"yearsNum":12.0,"tags":["geospatial","street-level","object detection","mapping","long-tail labels","world models","robotics","global"],"shortTitle":"Street-level object map"},{"ref":"MK-69833","title":"Screen-recorder company: engineering record and interaction traces","headline":"Screen-recorder startup: privacy-masked interaction traces plus full engineering record (~4.1K commits, ~430 design docs)","description":"Engineering record of a local-first screen recorder for macOS, Windows and Linux that captures video, audio, screenshots, keyboard/mouse, window and accessibility data with privacy masking at capture time and exports ML-ready interaction traces. Includes 3 private repos (~4,130 commits, ~820 PRs, ~220K lines of Python and Swift), ~430 design, research and decision docs, ~725 issues with in-app user feedback, and ~33 internal/QA recordings (~1.8 GB unique; design-tool tutorials and test runs) with synced video, input-event logs, window/accessibility metadata, transcripts and LLM task segments. Also sign-in, subscription and billing records and a few product-call transcripts. Customer recordings stay on users' machines and are not included.","category":"Agent traces","sector":"Software & internet","subsector":"Screen-recording software for ML interaction traces","dataTypes":["Screen recordings / UI traces","Video","Audio / speech","Code","Documents","Tabular / structured"],"useCases":["Computer-use & tool-use agents","Coding models","Vision & perception","SFT / instruction tuning"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"<10 GB","count":"~4,130 commits","volume":"~8 GB; 3 repos (~4,130 commits, ~820 PRs, ~220K LOC); ~430 design docs; ~725 issues","years":"2026","yearsOfHistory":"0.7","growing":"","systems":["Git","Issue tracker","Cloud storage","File storage","Auth","Billing","Team chat","Meeting recorder"],"contents":["Private repos (~4,130 commits, ~820 PRs, ~220K LOC)","Engineering docs: plans, research notes, decision records, runbooks (~430)","Issues with in-app user feedback (~725)","Internal/QA screen recordings (~33) with video and audio","Keyboard/mouse input-event logs (SQLite/JSONL)","Window and accessibility metadata","Recording transcripts and LLM task segments","Demo videos and product docs","Subscription and invoice records"],"formats":["Git","Markdown","MP4","SQLite","JSONL","CSV","DOCX","PDF"],"delivery":"Download","labelled":"Partly","labelTypes":["Transcripts","LLM task segments"],"personalData":"Low","deidentification":"","rights":"Via authorised intermediary","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"Screen-recording software","tools":[["chat","Team chat"],["tickets","Tickets"],["git","Git"],["calls","Calls"],["accounting","Accounting"],["files","File storage"]],"span":"8 months","volumeHead":"4.1K commits","sizeTB":0.008,"fromYear":2026,"toYear":2026,"yearsNum":0.7,"tags":["screen recorder","interaction traces","keyboard mouse events","accessibility tree","codebase","design docs","computer-use"],"shortTitle":"Screen-recorder startup archive"},{"ref":"MK-70164","title":"Analog and physical chip-design environment","headline":"RL environment for analog and physical chip design: 100 signoff-graded tasks with deterministic rewards and audited trajectories","description":"Containerised RL environment of 100 real chip-design tasks on two open process design kits: analog sizing, characterization, layout repair, routing, parasitic-aware repair, layout creation, mixed-signal integration and RTL-to-GDS. The agent works through a shell in a Linux workspace with the full open-source EDA toolchain and must ship a design passing signoff (SPICE specs across PVT corners, DRC, LVS, parasitic extraction); rewards in [0,1] with per-spec margins are computed by re-running real signoff tools in a separate container. Each task ships a reference solution and known-failing designs; generators create new layout-repair instances and private held-out variants exist for evaluation. CPU-only (4 vCPU / 8 GiB per episode), recorded and audited agent trajectories (788), token and log-prob capture for online RL.","category":"RL environment / benchmark","sector":"AI & data","subsector":"RL environment: semiconductor / EDA","dataTypes":["RL environment / simulator","Code","Agent trajectories"],"useCases":["RL environments","Evaluation / benchmarks","Coding models","Computer-use & tool-use agents"],"origin":"Expert-created","regions":["Europe"],"languages":["English"],"sizeBand":"","count":"100 tasks","volume":"100 tasks; 788 audited trajectories","years":"2026","yearsOfHistory":"","growing":"Yes","systems":["EDA tools","Docker containers","Git"],"contents":["100 chip-design task packages","Signoff graders (SPICE, DRC, LVS, parasitic extraction)","Reference solutions per task","Known-failing designs per task","Layout-repair instance generators","Private held-out variants","Recorded agent trajectories","Prebuilt Docker image and image recipe","Runbooks and per-task docs"],"formats":["Harbor task packages","Docker image","Python graders","GDS / SPICE / Verilog"],"delivery":"Buyer-hosted container","labelled":"Yes","labelTypes":["Reference solutions","Known-failing designs","Per-spec reward margins"],"personalData":"None","deidentification":"Not needed","rights":"Owner-held","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"RL environment: semiconductor / EDA","tools":[["docker","Docker"]],"span":"","volumeHead":"100 tasks","sizeTB":null,"fromYear":2026,"toYear":2026,"yearsNum":null,"tags":["rl environment","chip design","analog","eda","layout","verifiable rewards","agent benchmark","long-horizon"],"shortTitle":"Chip-design RL environment"},{"ref":"MK-71693","title":"US roofing and construction contractor: CRM, job costing, field ops","headline":"US roofing, painting and solar contractor: CRM, job costing, project images, roof reports, crews and telematics in one database","description":"Operating data of a US construction services contractor (residential and commercial roofing, painting, solar and maintenance), consolidated in one integrated database with cross-system links. Includes CRM contacts, deals, buildings, tickets, warranties, notes, emails, calls and SMS (~34K contacts, ~27K deals, 2022-2026); accounting with AP/AR invoices, vendors, contracts, job costs and ~790K GL entries; construction projects with estimates, budgets, change orders, logs, inspections, RFIs and ~170K image records; ~195K jobsite photo records with capture dates and geolocation; ~3.9K aerial roof measurement reports; dispatch tasks and field forms; crew and equipment schedules; payroll timecards (~32K entries); and vehicle telematics.","category":"Company operations archive","sector":"Construction & real estate","subsector":"Roofing, painting and solar contractor","dataTypes":["Tabular / structured","Image","Documents","Geospatial","Sensor / telemetry","Email"],"useCases":["Enterprise workflow agents","Vision & perception","Domain fine-tuning","Forecasting & time series","Retrieval / RAG"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"","count":"~790K GL entries","volume":"34,360 contacts; 27,457 deals; 791,886 GL entries; 171,945 project images + 195,582 jobsite photo records; 3,911 roof reports; 4,304 dispatch tasks; 32,483 time entries","years":"2022–2026","yearsOfHistory":"4.8","growing":"Yes","systems":["CRM","Accounting","Construction project management","Work management","HR & payroll","Field dispatch","Roof measurement reports","Jobsite photo app","Vehicle telematics","Database"],"contents":["CRM contacts, deals, warranties and tickets (~34K contacts, ~27K deals)","General-ledger entries (~790K) and job costs","Project estimates, budgets and change orders","Inspections, daily logs and RFIs","Project image records (~170K)","Geotagged jobsite photo records (~195K)","Roof measurement reports (~3.9K)","Dispatch tasks and field forms","Crew schedules and assignments","Timecards (~32K entries)","Vehicle positions and driving status"],"formats":["Database records (SQL)","Bulk exports and original files to confirm"],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"Owner-held","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"Mid-size company","employees":"51–200","industry":"Roofing, painting and solar contractor","tools":[["tickets","Tickets"],["git","Git"],["crm","CRM"],["database","Database"],["accounting","Accounting"],["hr","HR & payroll"]],"span":"5 years","volumeHead":"790K GL entries","sizeTB":null,"fromYear":2022,"toYear":2026,"yearsNum":4.8,"tags":["roofing","construction","contractor","job costing","jobsite photos","crm","field service","telematics","solar"],"shortTitle":"Roofing contractor operations"},{"ref":"MK-71756","title":"Auto hail-repair business: damage-to-payment insurance claim records","headline":"Auto hail-repair business: damage photos to insurer-approved estimates and payment, plus recorded field-sales calls","description":"Operating records of an insurance-funded automotive hail and paintless dent repair business. Job folders hold vehicle damage photos, scope sheets and the estimate chain (preliminary, insurer-approved, supplements, revisions and final paid estimates), insurer correspondence and payment evidence across ~3,260 repair orders; production and expense sheets tie operations to receipts, costs, margin and technician income. Also ~22,500 recorded field-sales interaction objects with transcripts and coaching, ~25K team-chat messages, ~33K business-phone call events and ~51K messages, work-phone group chats, CRM leads, a custom recruiting ATS, accounting, calendar and internal app code. Sources overlap; unique size and case linkage are not yet validated.","category":"Company operations archive","sector":"Automotive & transport","subsector":"Hail and paintless dent repair (insurance-funded)","dataTypes":["Image","Documents","Email","Audio / speech","Text","Chat & messaging","Tabular / structured","Code"],"useCases":["Enterprise workflow agents","Vision & perception","Domain fine-tuning","Speech & audio models","Computer-use & tool-use agents"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"100 GB-1 TB","count":"~3,260 repair orders","volume":"~0.5 TB gross (overlapping sources); ~3,260 repair orders; 22,500 sales-recording objects; ~25K chat messages; ~32.6K call events and 51K messages","years":"2020–2026","yearsOfHistory":"6","growing":"Yes","systems":["Email","File storage","Photo library","Spreadsheets","Shop management","Estimating","Team chat","Messaging","Telephony","Sales recording","CRM","ATS","Accounting","Git","Database"],"contents":["Vehicle damage photos","Scope sheets","Preliminary, insurer-approved, supplement and final estimates","Insurer correspondence","Payment evidence","Repair-order production and expense sheets","Recorded field-sales interactions with transcripts","Sales coaching objects","Business phone calls and texts","Team chat and work-phone group chats","CRM leads and appointments","Recruiting ATS records","Accounting attachments"],"formats":["PDF","JPG / photos","Google Sheets / XLSX","MBOX","Audio","JSON","SQL"],"delivery":"","labelled":"Partly","labelTypes":["Estimate revision chain to insurer approval and payment (linkage not validated)","Sales-call transcripts"],"personalData":"High","deidentification":"","rights":"Owner-held","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Hail and paintless dent repair","tools":[["chat","Team chat"],["git","Git"],["email","Email"],["crm","CRM"],["calls","Calls"],["database","Database"],["accounting","Accounting"],["design","Design files"],["sheet","Spreadsheets"],["shop","Commerce"],["hr","HR & payroll"],["files","File storage"]],"span":"6 years","volumeHead":"3.3K repair orders","sizeTB":0.52,"fromYear":2020,"toYear":2026,"yearsNum":6.0,"tags":["auto repair","hail damage","insurance claims","estimates","damage photos","sales calls","field sales","paintless dent repair"],"shortTitle":"Hail-repair insurance claims"},{"ref":"MK-74122","title":"Dental radiograph lesion-detection environment","headline":"Dental radiograph RL environment: 3,863 panoramic films, 6,645 dentist-labelled lesions, deterministic scorer and audited benchmark","description":"RL environment for periapical-lesion detection and Periapical Index (PAI) grading on 3,863 dental panoramic radiographs with 6,645 dentist-labelled lesions (bounding box plus grade 3-5). Includes a deterministic scorer with 24 CI tests, film-level train/validation/test splits with near-duplicate control, an audited seven-model benchmark (best model 0.19 vs a 0.83 two-dentist ceiling), a reward-hacking audit over 1,200 episodes and a post-training recipe. Films and labels are third-party open data (CC BY 4.0); the value sold is the environment, scorer and splits. Two sibling environments can be added as a suite: adult tooth charting and diagnosis (1,068 films) and paediatric disease detection (100 films).","category":"RL environment / benchmark","sector":"Healthcare & life sciences","subsector":"Dental imaging RL environment","dataTypes":["Image","RL environment / simulator","Tabular / structured"],"useCases":["RL environments","Evaluation / benchmarks","Vision & perception","Domain fine-tuning"],"origin":"Mixed","regions":["Southeast Asia"],"languages":["English"],"sizeBand":"<10 GB","count":"3,863 radiographs","volume":"~2 GB; 3,863 films; 6,645 dentist-labelled lesions","years":"Source films published 2024","yearsOfHistory":"","growing":"No","systems":["Hugging Face","Git"],"contents":["Panoramic dental radiographs","Lesion bounding boxes","PAI grades (3-5)","Film-level splits","Deterministic scorer","Seven-model audited benchmark","Inter-dentist expert ceiling","Post-training recipe","Sibling charting and paediatric environments"],"formats":["Parquet (JPEG + labels)","Python environment package"],"delivery":"Download","labelled":"Yes","labelTypes":["Lesion bounding boxes","PAI grades","Benchmark scores"],"personalData":"Low","deidentification":"Done by owner","rights":"Public sources (per-source licences)","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"Dental imaging RL environment","tools":[["git","Git"]],"span":"","volumeHead":"3.9K radiographs","sizeTB":0.002,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["dental","radiographs","medical imaging","rl environment","lesion detection","benchmark","vision"],"shortTitle":"Dental X-ray lesion environment"},{"ref":"MK-78006","title":"Indic two-speaker conversational speech","headline":"Indic two-speaker conversational speech: ~30K hours in 27 languages, 48 kHz speaker-isolated tracks, consented","description":"Permissioned five-minute, two-person remote conversations on self-chosen topics recorded across 29 Indian states and union territories, captured at 48 kHz with speaker-isolated tracks plus a mixed track (~1.1M tracks, 3 per conversation, ~30K hours, ~5 TB). Covers 27 languages including Hindi, Bengali, Marathi, Telugu, Tamil, Gujarati, Kannada, Malayalam, Odia, Punjabi, Urdu, Kashmiri, Maithili, Bhojpuri, Rajasthani and several under-resourced languages. Speaker metadata (age, gender, region, unique ID); human transcripts with diarisation available on request. Automated and human QC removes corrupted, incomplete or duplicate sessions; built for full-duplex speech-to-speech and TTS training, and scalable on request.","category":"Speech & conversation","sector":"AI & data","subsector":"Conversational speech data collection","dataTypes":["Audio / speech"],"useCases":["Speech & audio models","Pre-training","Domain fine-tuning"],"origin":"Purpose-collected (consented capture)","regions":["South Asia"],"languages":["Kashmiri","Hindi","Odia","Telugu","Dogri","Pashto","Marathi","Tamil","Urdu","Pahari","Punjabi","Sambalpuri","Bengali","Shina","Arabic","Malayalam","Gojri","Gujarati","Bhojpuri","Saurashtra","Maithili","Bihari","Kannada","Marwari","Nepali","Haryanvi","Rajasthani"],"sizeBand":"1-10 TB","count":"~30K hours","volume":"~4.99 TB; ~30K hours; 1.1M tracks","years":"2026","yearsOfHistory":"","growing":"Yes","systems":["Proprietary collection platform","Cloud storage"],"contents":["Two-speaker conversations","Speaker-isolated 48 kHz tracks","Mixed conversation tracks","Speaker metadata (age, gender, region, ID)","Session-level metadata","Transcripts with diarisation (on request)"],"formats":["OGG"],"delivery":"Cloud bucket (S3)","labelled":"Partly","labelTypes":["Speaker metadata","Transcripts and diarisation (on request)"],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"Conversational speech data collection","tools":[["files","File storage"]],"span":"","volumeHead":"30K hours","sizeTB":4.99,"fromYear":2026,"toYear":2026,"yearsNum":null,"tags":["speech","conversational audio","indic languages","full-duplex","tts","low-resource languages","india","speaker-isolated"],"shortTitle":"Indic conversational speech"},{"ref":"MK-82585","title":"Consented European consumer panel: search, browsing, video, social, shopping, AI-chat exports and mobile computer-use","headline":"Consented European consumer panel: ~7B records of search, browsing, video, social, shopping, AI-chat and mobile computer-use per person","description":"A paid, consented consumer panel of about 30,000 people, mostly in Germany, who share their full search, browser, video, maps, marketplace, social and professional-network history through official GDPR data-portability interfaces, refreshed weekly. Records are joined per person on one pseudonymous ID with age band, gender and region, giving cross-platform histories that include private-message exports, AI-assistant conversation exports, orders and receipts, and GPS for a subset. An Android on-device collector adds app sessions and computer-use traces (taps, scrolls, UI trees), and panellists can be recruited for commissioned collection such as dialect voice recordings and AI interviews. Delivered de-identified, only for purposes panellists opted into; AI training is a separate opt-in.","category":"Consumer behaviour","sector":"Marketing, media & publishing","subsector":"Consumer behaviour panel","dataTypes":["Tabular / structured","Screen recordings / UI traces","Chat & messaging","Text","Image","Video","Audio / speech"],"useCases":["Computer-use & tool-use agents","Pre-training","SFT / instruction tuning","Forecasting & time series","Evaluation / benchmarks","Speech & audio models"],"origin":"Purpose-collected (consented capture)","regions":["Europe"],"languages":["German","English"],"sizeBand":"1-10 TB","count":"~7B records","volume":"~7B behavioural records from ~30K panellists (~1.5 TB compressed); ~130M new records per month","years":"2006–2026","yearsOfHistory":"20","growing":"Yes","systems":["Panel collection platform","Mobile collector app"],"contents":["Search queries and clicks","Chrome browsing with dwell time","YouTube and TikTok watch histories","Instagram and Facebook content, follows and message exports","AI-assistant conversation exports","Marketplace orders, carts and reviews","Receipts from other merchants","Maps activity and GPS (subset)","Professional-network positions and skills (subset)","Android app sessions","Computer-use traces (taps, scrolls, UI trees)","Per-person demographics (age band, gender, region)"],"formats":["Parquet","CSV","JSON","JPG","PNG","MP4","WAV"],"delivery":"Download; Cloud bucket (S3)","labelled":"Yes","labelTypes":["Intent and topic categories"],"personalData":"High","deidentification":"Done by owner","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"Consumer behaviour panel","tools":[],"span":"20 years","volumeHead":"7B records","sizeTB":3.0,"fromYear":2006,"toYear":2026,"yearsNum":20.0,"tags":["consumer panel","browsing history","search","clickstream","social media","ai chat logs","shopping","computer use","ui traces","gdpr","germany"],"shortTitle":"European consumer data panel"},{"ref":"MK-85194","title":"Middle East developer and general contractor: 20+ years, ~23,700 files incl. ~6,500 native DWG drawings, BOQs, tenders and job costing","headline":"Developer/contractor archive: ~6,500 AutoCAD drawings (architectural, structural, MEP, as-built, revisions), BOQs, tenders, contracts, site reports, job costing, GL, invoices and POs","description":"Project and company archive of a developer and general contractor with 20+ years of work: native AutoCAD drawings across architectural, structural and MEP disciplines including as-builts and revision history, bills of quantities, tenders, contracts, site reports, a kitchen-design archive, project cost tracking, general ledger, invoices and purchase orders, and corporate documents.","category":"Company operations archive","sector":"Real estate & construction","subsector":"Development and contracting","dataTypes":["CAD / 3D","Documents","Tabular / structured"],"useCases":["Engineering agents","Enterprise workflow agents","Document understanding"],"origin":"First-party company records","regions":["Middle East"],"languages":[],"sizeBand":"10-100 GB","count":"~23,700 files","volume":"~23,700 files; 25 GB; ~8,200 PDFs; ~6,500 DWG; ~2,500 Word; ~1,400 Excel","years":"~15 years of files","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":["DWG, PDF, DOCX, XLSX"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Medium","deidentification":"","rights":"Owner-held","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"Mid-size company","employees":"51–200","industry":"Development and contracting","tools":[["sheet","Excel"],["autocad","AutoCAD"]],"span":"15 years","volumeHead":"24K files","sizeTB":0.025,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["construction","cad","dwg","boq","tenders","job costing"],"shortTitle":"Gulf contractor drawings archive"},{"ref":"MK-85212","title":"DeFi protocol with formal proofs","headline":"DeFi protocol codebase: Solidity contracts with Lean formal proofs, ~160 test files, CI and a bug log","description":"A young but heavily tested DeFi protocol from about two months of intensive 2026 development: Solidity contracts with Lean formal proofs, a TypeScript web app, agent-oriented documentation, a bug-collection log and demos. 7 private repos with 675 commits, 180 PRs, ~160 test files and 13 CI workflows; no overlap with public code found. Code only. Offered by an authorised reseller.","category":"Code & software","sector":"Crypto & web3","subsector":"DeFi protocol","dataTypes":["Code","Documents"],"useCases":["Coding models","Evaluation / benchmarks"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"<10 GB","count":"675 commits","volume":"32 MB code; 7 repos; 675 commits; 180 PRs; ~160 test files","years":"2026","yearsOfHistory":"0.2","growing":"","systems":["Git"],"contents":["Solidity protocol contracts","Lean formal proofs","TypeScript web app","Agent-oriented documentation","Bug-collection log","Test files (~160)","CI workflows (13)","Pull requests with history (180)"],"formats":["Git bundle"],"delivery":"Download","labelled":"Partly","labelTypes":["Formal proofs","Tests","Bug log"],"personalData":"None","deidentification":"Not needed","rights":"Via authorised intermediary","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"DeFi protocol","tools":[["git","Git"]],"span":"2 months","volumeHead":"675 commits","sizeTB":3e-05,"fromYear":2026,"toYear":2026,"yearsNum":0.2,"tags":["defi","solidity","formal verification","lean","smart contracts","codebase","tests"],"shortTitle":"DeFi protocol with proofs"},{"ref":"MK-86359","title":"US tax-return computation cases: source documents to correct federal and state returns","headline":"US tax-return computation cases: thousands of expert-built source-document sets with correctly computed federal and state returns","description":"Thousands of cases, each pairing the documents needed to file a US individual tax return (W-2s, 1099s and other forms, dozens per case) with the correctly computed federal and state return as ground truth. Cases were built by tax-software subject-matter experts and cover 2022-2026. Intended for training and evaluating models on tax calculation, filing and financial question answering.","category":"Financial & transactional","sector":"Financial services & insurance","subsector":"Tax filing software","dataTypes":["Documents","Tabular / structured"],"useCases":["Evaluation / benchmarks","Domain fine-tuning","SFT / instruction tuning","RL environments"],"origin":"Expert-created","regions":["North America"],"languages":[],"sizeBand":"","count":"Thousands of cases","volume":"Thousands of cases, dozens of documents each","years":"2022–2026","yearsOfHistory":"4","growing":"","systems":[],"contents":["Tax source documents (W-2s, 1099s, other forms)","Correct federal returns","Correct state returns","Case-level input/output pairs"],"formats":["PDF","JSON","XML"],"delivery":"","labelled":"Yes","labelTypes":["Correct computed returns (ground truth)"],"personalData":"None","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"Tax filing software","tools":[],"span":"4 years","volumeHead":"Thousands of cases","sizeTB":null,"fromYear":2022,"toYear":2026,"yearsNum":4.0,"tags":["tax","tax returns","benchmark","financial reasoning","expert-built","w-2","1099","evaluation"],"shortTitle":"US tax-return cases"},{"ref":"MK-86382","title":"DTC consumer brand ($20M+ revenue): 6-year full operating record","headline":"DTC pet-supplement brand ($20M+ revenue): 6 years of SOPs, email, support tickets, ads and orders","description":"The full six-year operating record of a direct-to-consumer pet-supplement brand that grew from launch to $20M+ revenue with ~20 people and then restructured: ~1.27 TB of shared-drive files (SOPs, financial models, P&Ls, budgets, product and creative briefs, supplier and agency documents, decks, ad creative and product media), ~30 GB of operator email with suppliers, manufacturers, 3PLs, agencies and retailers, knowledge-base wikis, ~100K support tickets linked to orders, email/SMS campaigns and A/B tests, paid-ads history, ~524K order and subscription records, ~100 code repositories and accounting records. Customer PII, formulation IP and legal material are excluded.","category":"Company operations archive","sector":"Retail & e-commerce","subsector":"DTC consumer brand (pet supplements)","dataTypes":["Email","Documents","Tabular / structured","Image","Video","Code","Chat & messaging"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG","Pre-training"],"origin":"First-party company records","regions":["North America"],"languages":["English"],"sizeBand":"1-10 TB","count":"~1.4 TB; ~100K tickets","volume":"~1.4 TB; ~20K documents; ~100K support tickets; ~524K order/subscription records; ~100 repos","years":"2020–2026","yearsOfHistory":"6","growing":"Yes","systems":["Email","File storage","Help desk","E-commerce","Marketing automation","Git","Accounting"],"contents":["Operator email with suppliers and 3PLs","SOPs","Financial models and P&Ls","Support tickets linked to orders","Email/SMS campaigns and A/B tests","Paid-ads history","Order and subscription records","Ad creative library","Code repositories"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"Done by owner","rights":"Owner-held","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"DTC consumer brand","tools":[["tickets","Tickets"],["git","Git"],["email","Email"],["accounting","Accounting"],["shop","Commerce"],["analytics","Analytics & marketing"],["files","File storage"]],"span":"6 years","volumeHead":"1.4 TB","sizeTB":1.4,"fromYear":2020,"toYear":2026,"yearsNum":6.0,"tags":["dtc","ecommerce","consumer brand","support tickets","marketing","sops","shopify"],"shortTitle":"DTC brand operating record"},{"ref":"MK-86791","title":"Large Latin American construction firm: ~250K BIM/CAD files from ~390 real projects","headline":"Large Latin American construction firm: ~250K BIM/CAD files from ~390 real projects with versions, RFIs and clashes","description":"Project records of a large Latin American construction company, ~200 million sq ft of BIM and CAD coverage across ~390 projects (~265 buildings, ~120 urban developments, 4 shopping centres and a cultural centre). Revit models (~25K), DWG drawings (~7.5K), Navisworks coordination models, Civil 3D/DGN files (~4.4K) and ~210K DWFX/PDF/XLSX documents across architecture, structure, plumbing, lighting, electrical and life safety. Versions, RFIs, clash reviews and approvals link problems to their corrections, and calculations and specifications give testable constraints; itemized costs are included, and issue-to-RFI-to-clash-correction sequences can be reconstructed for ~29 projects. Each project ships with discipline folders, a machine-readable file index (file, discipline, document type, format) and a project data sheet (units, codes, dates, site).","category":"Industrial & sensor","sector":"Construction & real estate","subsector":"General contractor / BIM and CAD project archive","dataTypes":["3D / CAD","Documents","Tabular / structured"],"useCases":["Domain fine-tuning","Enterprise workflow agents","Computer-use & tool-use agents","Evaluation / benchmarks","Pre-training","World models"],"origin":"First-party company records","regions":["Latin America"],"languages":[],"sizeBand":"","count":"~250K files across ~390 projects","volume":"~250K files across ~390 projects (~25K RVT; ~7.5K DWG; ~350 NWD; ~4.4K Civil 3D; ~210K DWFX/PDF/XLSX)","years":"2009–2026","yearsOfHistory":"17","growing":"","systems":["Design files","File storage","Construction document management"],"contents":["Revit BIM models","DWG 2D drawings","Navisworks coordination models","Civil 3D / DGN urban-development files","DWFX/PDF/XLSX project documentation","Model version history","RFIs and clash reviews linked to corrections","Calculations, specifications and approvals","Itemized costs","Per-project file index and project data sheet"],"formats":["RVT","DWG","NWD","Civil 3D / DGN","DWFX","PDF","XLSX"],"delivery":"API","labelled":"No","labelTypes":[],"personalData":"None","deidentification":"Not needed","rights":"Via authorised intermediary","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"General contractor / BIM and CAD project archive","tools":[["design","Revit"],["files","File storage"]],"span":"17 years","volumeHead":"250K files across ~390 projects","sizeTB":null,"fromYear":2009,"toYear":2026,"yearsNum":17.0,"tags":["construction","bim","cad","revit","navisworks","rfi","clash detection","latin america","engineering history"],"shortTitle":"LatAm construction BIM library"},{"ref":"MK-89189","title":"Retrosynthesis RL data with verifiers: 100K generated chemistry records","headline":"100K generated retrosynthesis and drug-discovery training records with verifiers and a RL environment, built from public patent routes and proprietary molecular data","description":"Chemistry training data for retrosynthesis and drug discovery: 100K records generated by an expert-guided method from public patent routes and proprietary molecular data, each automatically verified (a small share chemist-reviewed), with verifiers and a RL environment. Never sold before.","category":"RL environment / benchmark","sector":"Healthcare & life sciences","subsector":"Chemistry / drug discovery","dataTypes":["Text","RL environment / simulator"],"useCases":["RL environments","Domain fine-tuning","SFT / instruction tuning"],"origin":"Generated / simulated","regions":["Not stated"],"languages":[],"sizeBand":"","count":"100K records","volume":"100K records (JSON)","years":"Routes 2011-2025","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":["JSON"],"delivery":"","labelled":"Yes","labelTypes":[],"personalData":"None","deidentification":"Not needed","rights":"Owner-held","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"Chemistry / drug discovery","tools":[],"span":"","volumeHead":"100K records","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["chemistry","retrosynthesis","drug discovery","verifiers","rl"],"shortTitle":"Retrosynthesis RL with verifiers"},{"ref":"MK-89632","title":"Bonded warehouse and trading company: ERP and email","headline":"Bonded warehouse and trading company: 14 years of ERP (GL, A/R, A/P, inventory) and 17 years of email, Arabic/English","description":"Records of a customs-controlled bonded warehouse that stores imported goods under customs supervision before duties are paid, alongside a general trading business. The ERP database (since 2012) covers chart of accounts, journal entries, general ledger, bank transactions and reconciliations, monthly income statements per entity, customer and supplier masters, sales and supplier invoices, credit notes, receipts, payments and ageing, purchase orders, and item master, stock movements, levels and valuation. Company email (since 2009, on-premises plus hosted) holds correspondence with warehouse clients, customers and suppliers, POs, invoices and quotations, customs and clearance paperwork, contracts, onboarding and service issues, and internal management threads, in Arabic and English.","category":"Company operations archive","sector":"Logistics, distribution & wholesale","subsector":"Bonded warehousing and general trading","dataTypes":["Tabular / structured","Email","Documents"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Forecasting & time series","Retrieval / RAG"],"origin":"First-party company records","regions":["Middle East"],"languages":["Arabic","English"],"sizeBand":"10-100 GB","count":"","volume":"33 GB (email 18 GB; ERP database 15 GB)","years":"2009–2026","yearsOfHistory":"17","growing":"Yes","systems":["ERP","Email"],"contents":["General ledger and journal entries","Bank transactions and reconciliations","Monthly income statements per entity","Customer master, sales invoices and receivables ageing","Supplier master, purchase orders and payables ageing","Item master, stock movements and valuation","Customs and clearance paperwork","Client and supplier correspondence","Contracts"],"formats":["SQL Server extract","CSV","PST"],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Ready now","stage":"Inventory ready","scale":"Mid-size company","employees":"51–200","industry":"Bonded warehousing and general trading","tools":[["email","Email"],["erp","ERP"]],"span":"17 years","volumeHead":"33 GB","sizeTB":0.033,"fromYear":2009,"toYear":2026,"yearsNum":17.0,"tags":["bonded warehouse","customs","trading","erp","accounting","email archive","arabic","inventory"],"shortTitle":"Bonded warehouse ERP and email"},{"ref":"MK-91141","title":"Physics-graded CAD design RL environment (load-bearing brackets)","headline":"Code-writing agents design load-bearing brackets in CAD, graded by finite-element physics against an optimised reference: 46-task benchmark plus unlimited seeded tasks","description":"Agents write CAD code to design aluminium brackets that carry real loads; designs are graded by finite-element simulation against a topology-optimised reference (stiffness, strength, mass). Frozen 46-task benchmark in 4 difficulty levels plus unlimited procedurally generated tasks; Gymnasium and Python APIs, HTTP grading server and Docker. Frontier models produce feasible designs on 35-50% of tasks and none come within 10% of optimal mass.","category":"RL environment / benchmark","sector":"Manufacturing","subsector":"Mechanical design","dataTypes":["RL environment / simulator","3D / CAD","Code"],"useCases":["RL environments","Evaluation / benchmarks","Coding models"],"origin":"Generated / simulated","regions":["Not stated"],"languages":[],"sizeBand":"","count":"46 tasks + procedural","volume":"46-task benchmark + unlimited procedural tasks","years":"2026","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":["Python package","Docker image"],"delivery":"","labelled":"Yes","labelTypes":[],"personalData":"None","deidentification":"Not needed","rights":"Owner-held","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"Mechanical design","tools":[["docker","Docker"]],"span":"","volumeHead":"46 tasks + procedural","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["cad","fea","rl environment","mechanical engineering","benchmark"],"shortTitle":"Physics-graded CAD environment"},{"ref":"MK-91733","title":"Crypto exchange company: 5-year archive","headline":"Decentralised perpetual-exchange company: 5 years of trading data for 200K+ users, 160+ repos, support and ops records","description":"Five years of records from a Web3 company that built decentralised perpetual exchanges, smart contracts and trading infrastructure: historical perpetual-exchange trading data for 200K+ users; 167 code repositories with ~48K pull requests across backend (Node/TypeScript, Java, Python, Go, Rust), web, mobile and Solidity; exchange smart contracts with security audit reports; market-making algorithms and data; operations knowledge bases for low-latency on-chain exchanges and blockchain networks; token listing, deployment, vesting and liquidity-provision records; plus product telemetry, support tickets, design files, docs, finance and marketing data and email. ~0.38 TB in total, of which ~0.12 TB is proprietary exchange and trading material.","category":"Company operations archive","sector":"Crypto & web3","subsector":"Decentralised derivatives exchange","dataTypes":["Tabular / structured","Code","Documents","Email","Chat & messaging","Image","Sensor / telemetry"],"useCases":["Coding models","Forecasting & time series","Domain fine-tuning","Enterprise workflow agents","Retrieval / RAG"],"origin":"First-party company records","regions":["Southeast Asia"],"languages":[],"sizeBand":"100 GB-1 TB","count":"200K+ users; 167 repos","volume":"~0.38 TB (~0.12 TB proprietary trading and exchange assets); 200K+ users; 167 repos, ~48K PRs","years":"2021–2026","yearsOfHistory":"5","growing":"","systems":["Database","Git","Email","File storage","Spreadsheets","Wiki","Customer support","Analytics","Design files","Marketing","Cloud storage","Hosting"],"contents":["Exchange trading and transaction records (200K+ users)","Product telemetry","Market-making algorithms and data","Exchange and smart-contract code (160+ repos, ~48K PRs)","Smart-contract security audit reports","Node and private/public chain operation guides","Customer support conversations","Internal knowledge base and task management","Product designs and workflow diagrams","Invoices, fundraise and finance sheets","Ads and marketing campaign data","Token deployment, vesting and liquidity materials"],"formats":["JSONL","PDF","JPG","ZIP (Git)"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"Done by owner","rights":"Owner-held","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Decentralised derivatives exchange","tools":[["tickets","Tickets"],["git","Git"],["email","Email"],["wiki","Docs & wiki"],["database","Database"],["design","Design files"],["sheet","Spreadsheets"],["analytics","Analytics & marketing"],["files","File storage"]],"span":"5 years","volumeHead":"200K users","sizeTB":0.38,"fromYear":2021,"toYear":2026,"yearsNum":5.0,"tags":["crypto exchange","perpetuals","trading data","defi","smart contracts","codebase","market making","web3"],"shortTitle":"Crypto exchange archive"},{"ref":"MK-95041","title":"Egocentric factory video: bag manufacturing","headline":"Egocentric 1080p video of real bag-factory work (sewing, binding, assembly) in South Asia, 1,000+ hours, capture to spec","description":"First-person video of real workers doing their normal jobs on a bag-manufacturing floor in Nepal: sewing, binding, trimming and assembling bags on industrial sewing machines, shot on head-mounted phones with an ultrawide lens so both hands and the work surface stay in frame through long continuous sessions (up to ~1 hour each). Covers dense two-handed handling of soft, deformable materials (fabric, straps, zippers) and industrial machine operation in a cluttered real factory. 1,000+ hours from the existing library plus new capture to specification; an on-site team can scale capture to other manufacturing, agriculture and construction tasks. Raw RGB video today (no depth or IMU); task/activity labels, action timestamps, scene and device metadata and anonymised worker IDs on request.","category":"Physical-world video","sector":"Manufacturing","subsector":"Bag / textile manufacturing (egocentric capture)","dataTypes":["Video","Audio / speech"],"useCases":["Robotics / embodied AI","World models","Vision & perception"],"origin":"Purpose-collected (consented capture)","regions":["South Asia"],"languages":["Nepali"],"sizeBand":"1-10 TB","count":"1,000+ hours","volume":"1,000+ hours (existing plus capture to specification); ~3.4 GB/hour","years":"2026","yearsOfHistory":"","growing":"Yes","systems":["Head-mounted phone cameras","Cloud storage"],"contents":["Egocentric sewing and binding video","Bag assembly sessions","Deformable-material manipulation","Industrial sewing-machine operation","Incidental ambient audio (removable)","Per-session files up to ~1 hour"],"formats":["MP4 (H.264 1080p 29.97 fps)","AAC"],"delivery":"Cloud bucket (S3); Cloud bucket (GCS); Download","labelled":"No","labelTypes":[],"personalData":"Low","deidentification":"Planned / on request","rights":"Owner-held","sample":"Available","readiness":"Ready now","stage":"Inventory ready","scale":"","employees":"","industry":"Bag / textile manufacturing","tools":[["files","File storage"]],"span":"","volumeHead":"1K hours","sizeTB":3.4,"fromYear":2026,"toYear":2026,"yearsNum":null,"tags":["egocentric video","factory","sewing","deformable objects","manipulation","robotics","vla","nepal"],"shortTitle":"Bag factory egocentric video"},{"ref":"MK-12348","title":"Established US landscaping company: 7 years of jobs, routes, invoices and job photos","headline":"US landscaping company: 7 years of field-service jobs, crew routes, invoices, payments and ~9,000 job emails with photos","description":"Operating records of a sizeable US landscaping company (mowing, lawn care, mulching and more) run through a field-service CRM: invoices, purchase orders, crew routes and payments over seven years, plus ~9,000 emails with invoices, job photos and job descriptions, and company files in Drive. Owner states ~10 TB total.","category":"Company operations archive","sector":"Consumer services","subsector":"Landscaping and lawn care","dataTypes":["Tabular / structured","Email","Image","Documents"],"useCases":["Enterprise workflow agents","Vision & perception","Forecasting & time series","Domain fine-tuning"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"10-100 TB","count":"~9,000 job emails","volume":"~10 TB; ~9,000 job emails","years":"2019–2026","yearsOfHistory":"7","growing":"Yes","systems":["Field service CRM","Email","File storage"],"contents":["Invoices","Purchase orders","Crew routes","Payments","Job photos","Job descriptions","Job emails (~9,000)","Company files"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Landscaping and lawn care","tools":[["email","Email"],["crm","CRM"],["files","File storage"]],"span":"7 years","volumeHead":"9K job emails","sizeTB":10.0,"fromYear":2019,"toYear":2026,"yearsNum":7.0,"tags":["landscaping","field service","routes","invoices","job photos","lawn care","home services"],"shortTitle":"Landscaping company records"},{"ref":"MK-13381","title":"AI-focused marketing agency: complete archive","headline":"Closed marketing agency for AI companies: email, docs, ad creatives, ~1.1 TB finished video and call recordings","description":"Full retained archive of a closed marketing agency whose clients were mostly fast-growing AI companies. Includes workspace email, Drive documents, financial spreadsheets, campaign strategies and ad creatives (~0.5 TB), ~1.1 TB of finished videos, ~80 GB of recorded stand-ups, sales calls and strategy meetings, social-post history with performance stats (billions of cumulative views), project planning, payroll records for 100+ contractors and employees, calendar, billing records and team chat.","category":"Company operations archive","sector":"Marketing, media & publishing","subsector":"Marketing agency (AI-company clients)","dataTypes":["Email","Documents","Video","Audio / speech","Image","Chat & messaging","Tabular / structured","Text"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Speech & audio models","Retrieval / RAG","Pre-training"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"1-10 TB","count":"~1.1 TB finished video","volume":"~1.7 TB total (documents ~0.5 TB, video ~1.1 TB, call recordings ~80 GB)","years":"","yearsOfHistory":"","growing":"No","systems":["Email","File storage","Video review","Meeting recorder","Social scheduling","Project management","Payroll","Calendar","Billing","Team chat"],"contents":["Client and internal email","Campaign strategies","Ad creatives","Finished marketing videos (~1.1 TB)","Recorded stand-ups, sales and strategy calls","Social posts with performance stats","Financial spreadsheets","Project planning tasks","Payroll records","Billing records","Team chat"],"formats":["Platform exports"],"delivery":"","labelled":"Partly","labelTypes":["Social performance metrics"],"personalData":"High","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Marketing agency","tools":[["chat","Team chat"],["tickets","Tickets"],["email","Email"],["calls","Calls"],["accounting","Accounting"],["hr","HR & payroll"],["calendar","Calendar"],["files","File storage"]],"span":"","volumeHead":"1.1 TB finished video","sizeTB":1.7,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["marketing agency","ad creatives","video","sales calls","social media","campaigns","payroll","closed company"],"shortTitle":"AI marketing agency archive"},{"ref":"MK-15802","title":"Aircraft MRO lifecycle records","headline":"Aircraft cabin-equipment MRO lifecycle records: ~1M repair, status, parts and flight-linked events in 7 relational tables","description":"Repair lifecycle of airline cabin equipment (galley carts/trolleys and related units) handled by a maintenance, repair and overhaul (MRO) operation across 6 stations, captured on the owner's repair-tracking platform (part migrated from earlier records). Seven core tables: asset registry, repair orders (scope of work, work done, status), repair status history (status changes with actor and JSON diffs), work items (parts and tasks, quantities, completion), carts (type, state, station, MRO due date), flights (flight number, route, schedule, aircraft type, status) and cart state events (scan or manual state transitions linked to flights), plus an optional audit-events table. Names, emails, organisation names, invoice and bank details and some pricing are removed; serial, part and reference numbers and station codes are tokenised; repair types, scopes, statuses, timestamps and outcomes are kept. For predictive maintenance, turnaround and logistics models and agents that run MRO workflows.","category":"Industrial & sensor","sector":"Automotive & transport","subsector":"Aviation MRO (cabin equipment repair)","dataTypes":["Tabular / structured"],"useCases":["Enterprise workflow agents","Forecasting & time series","Domain fine-tuning","Computer-use & tool-use agents"],"origin":"First-party company records","regions":["South Asia"],"languages":[],"sizeBand":"","count":"~1M lifecycle records","volume":"~1M lifecycle records in 7 core tables (~1.2M with audit events)","years":"2022–2026","yearsOfHistory":"4","growing":"Yes","systems":["Repair-tracking platform","Database"],"contents":["Asset registry and lifecycle status","Repair orders with scope and work done","Repair status history with actor and JSON diffs","Work items: parts, tasks, quantities","Cart states, stations and MRO due dates","Flight records (route, schedule, aircraft type)","Scan-based cart state events linked to flights","Optional audit-events table"],"formats":["Relational tables"],"delivery":"","labelled":"Yes","labelTypes":["Repair statuses and outcomes","Scopes of work"],"personalData":"Low","deidentification":"Done by owner","rights":"Owner-held, conditions to confirm","sample":"On request / under NDA","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"Aviation MRO","tools":[["database","Database"]],"span":"4 years","volumeHead":"1M lifecycle records","sizeTB":null,"fromYear":2022,"toYear":2026,"yearsNum":4.0,"tags":["aviation","mro","maintenance records","galley carts","repair orders","predictive maintenance","workflow logs","relational"],"shortTitle":"Aircraft maintenance records"},{"ref":"MK-15872","title":"Smart-contract wallet: contracts, proofs and SDK","headline":"ERC-4337 smart-contract wallet: Solidity with Lean proofs, SDK, bundler/paymaster and recovery services (55 private repos)","description":"Full smart-contract wallet stack built 2022-2026: ERC-4337 wallet contracts in Solidity with Lean formal proofs, a TypeScript wallet SDK, bundler and sponsor (paymaster) gateway, transaction and op-sync APIs, social-recovery and email-guardian services, a KMS signer, push notifications, an asset monitor, an email-proof prover service, and web and browser-extension wallets. 55 private repos with ~3,600 commits (about a third also in earlier public releases), ~240 PRs and ~200 test files. Team chat, docs/sheets, product specs, roadmaps, engineering docs and meeting notes are also offered. Offered by an authorised reseller.","category":"Code & software","sector":"Crypto & web3","subsector":"Smart-contract wallet","dataTypes":["Code","Documents","Chat & messaging"],"useCases":["Coding models","Evaluation / benchmarks","Domain fine-tuning"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"","count":"~3,600 commits","volume":"50+ private repos / ~3.6K commits / ~240 PRs; documentation est. 200-3,000 pages","years":"2022–2026","yearsOfHistory":"4","growing":"","systems":["Git","Team chat","Wiki"],"contents":["Solidity wallet contracts","Lean formal proofs","TypeScript wallet SDK","Bundler and paymaster back-ends","Social-recovery and email-guardian services","KMS signer service","Web and browser-extension wallets","Test files (~200)","Product specs and roadmaps","Meeting notes and team chat"],"formats":["Git bundle","DOCX","XLSX","HTML / Markdown","PDF"],"delivery":"Download","labelled":"Partly","labelTypes":["Formal proofs","Tests"],"personalData":"Low","deidentification":"","rights":"Via authorised intermediary","sample":"","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"Smart-contract wallet","tools":[["chat","Team chat"],["git","Git"],["wiki","Docs & wiki"]],"span":"4 years","volumeHead":"3.6K commits","sizeTB":null,"fromYear":2022,"toYear":2026,"yearsNum":4.0,"tags":["smart contracts","solidity","formal verification","lean","account abstraction","wallet","typescript","crypto"],"shortTitle":"Smart-contract wallet codebase"},{"ref":"MK-15903","title":"UK mortgage-origination software as a computer-use RL environment, plus ~2,000 underwriter-reviewed case tasks","headline":"A live mortgage-origination SaaS platform offered as a computer-use RL environment, with ~1,000-2,000 task/rubric pairs built from ~2,000 anonymised historic cases, each reviewed twice by human underwriters","description":"A UK mortgage-origination platform used by lenders, offered as a RL environment for computer-use agents, together with ~2,000 anonymised historic mortgage cases packaged as task/rubric pairs and double-reviewed by human underwriters. Environment and case set only, no general company corpus.","category":"RL environment / benchmark","sector":"Financial services","subsector":"Mortgage lending","dataTypes":["RL environment / simulator","Tasks & rubrics"],"useCases":["RL environments","Computer-use & tool-use agents","Finance agents"],"origin":"First-party company records","regions":["Europe"],"languages":[],"sizeBand":"","count":"~2,000 cases","volume":"1 environment; ~2,000 cases; ~1,000-2,000 task/rubric pairs","years":"~7 years operating","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"Yes","labelTypes":[],"personalData":"Low","deidentification":"Anonymised","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Mid-size company","employees":"51–200","industry":"Mortgage lending","tools":[],"span":"7 years","volumeHead":"2K cases","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["rl environment","computer use","mortgage","underwriting","rubrics"],"shortTitle":"Mortgage underwriting environment"},{"ref":"MK-16287","title":"Enterprise call-centre audio with outcomes and coaching labels","headline":"Dual-channel B2B call-centre audio with transcripts, outcomes and expert coaching labels: ~100K hours, 2022-2026","description":"Live, unscripted dual-channel phone calls (rep and customer on separate tracks) from 15 enterprise organisations in auto lending and collections, higher-education enrolment, insurance, home services and energy, banking and real estate; US locale across 40+ states, ~25% bilingual Spanish/English with code-switching, inbound/outbound roughly 50/50. ASR transcripts with millisecond utterance timestamps and speaker labels; outcome codes (payment committed, refused, escalated, repossession, enrollment, closed-won/lost); ~112K expert utterance-level coaching annotations; ~95K objection-to-outcome pairs; rep performance scores; and, for collections, per-call financial metadata (days delinquent, balances, delinquency bucket). ~9.5K reps, the same speakers tracked across calls for 2+ years, and a 700K+ calls/month active pipeline.","category":"Speech & conversation","sector":"Financial services & insurance","subsector":"Contact-centre sales and collections calls","dataTypes":["Audio / speech","Text","Tabular / structured"],"useCases":["Speech & audio models","Domain fine-tuning","Preference / RLHF","Evaluation / benchmarks","Enterprise workflow agents"],"origin":"First-party company records","regions":["North America"],"languages":["English","Spanish"],"sizeBand":"1-10 TB","count":"~100K hours (~4M calls, estimated)","volume":"3.1 TB; 100,000+ hours (estimated); ~4M calls; 700K+ calls/month pipeline; 112K human annotations; 95K objection-to-outcome pairs","years":"2022–2026","yearsOfHistory":"4","growing":"Yes","systems":["Telephony","Contact-centre dialers","Call analytics / coaching platform"],"contents":["Dual-channel call recordings","ASR transcripts with utterance timestamps","Call outcome codes","Expert utterance-level coaching annotations","Objection-to-outcome pairs","Rep performance scores","Collections financial metadata and delinquency buckets","AI-generated coaching suggestions (separate)"],"formats":["WAV","MP4","JSON","CSV"],"delivery":"Cloud bucket (GCS); Cloud bucket (S3)","labelled":"Yes","labelTypes":["Utterance-level coaching labels","Win/loss outcome codes","Rep performance scores","Delinquency buckets"],"personalData":"High","deidentification":"","rights":"To be confirmed","sample":"On request / under NDA","readiness":"On request","stage":"Inventory ready","scale":"","employees":"","industry":"Contact-centre sales and collections calls","tools":[["calls","Calls"]],"span":"4 years","volumeHead":"100K hours","sizeTB":3.1,"fromYear":2022,"toYear":2026,"yearsNum":4.0,"tags":["call centre","speech","dual-channel audio","collections","sales calls","bilingual","transcripts","coaching labels","outcomes"],"shortTitle":"Call-centre audio with coaching"},{"ref":"MK-16965","title":"Consumer life app: 52M trip-grade GPS pings with activity labels, plus app event log","headline":"52M production GPS pings at ~3 s cadence and ~4 m median accuracy, with speed, heading and activity on every ping, from 10k app users over 19 months","description":"Production location telemetry from a consumer life-organiser app (map, calendar and assistant), collected in the background on users' phones from 2024 to 2026. About 52M GPS pings from roughly 10k users and 63k user-days, an estimated 126k trips, with about two thirds of pings in a vehicle. Each ping carries coordinates, horizontal and vertical accuracy, altitude, speed, heading, on-device activity type and confidence (vehicle, foot, bicycle, running, still), motion start and stop events, odometer and battery state, plus precomputed H3 cells. A core longitudinal panel of about 330 users was active 30+ days. A separate log of 1.56M app events (76 event types) covers map, calendar and assistant use. Dev and test data are excluded. User IDs are re-keyed, names, emails and free text are removed, and trip ends near home and work are trimmed. Location consent and terms review are in progress.","category":"Industrial & sensor","sector":"Software & internet","subsector":"Consumer life-organiser app (map, calendar, assistant)","dataTypes":["Location / GPS","Sensor data","Event logs"],"useCases":["World models","Trajectory models","Mode detection","Map-matching and ETA","Telematics and driving risk","Agent planning"],"origin":"Platform / user-generated","regions":["North America","Europe","Oceania"],"languages":[],"sizeBand":"10-100 GB","count":"52M GPS pings","volume":"~52M GPS pings (6.5 GB Parquet, ~61 GB uncompressed), ~10k users, 63k user-days, ~126k trips; 1.56M app events, 76 event types","years":"2024–2026","yearsOfHistory":"2","growing":"No","systems":["Background geolocation SDK","Google BigQuery","Mixpanel"],"contents":["GPS pings with accuracy, speed, heading and altitude","Activity type and confidence per ping","Motion start/stop events","H3 cells","App event log (map, calendar, assistant)"],"formats":["Parquet","JSONL","CSV"],"delivery":"Parquet or CSV, IDs re-keyed across tables, with data dictionary","labelled":"Yes","labelTypes":["Activity type and confidence per ping","Motion events","App event types"],"personalData":"High","deidentification":"IDs re-keyed; names, emails and free text removed; home/work trip ends trimmed; coarse H3 only","rights":"To be confirmed","sample":"Available on request","readiness":"On request","stage":"Inventory ready","scale":"","employees":"","industry":"Consumer life-organiser app","tools":[["mixpanel","Mixpanel"]],"span":"2 years","volumeHead":"52M GPS pings","sizeTB":0.06,"fromYear":2024,"toYear":2026,"yearsNum":2.0,"tags":["gps","telematics","trajectories","mobility","activity recognition","mode detection","location","consumer app"],"shortTitle":"Trip-grade phone GPS with activity labels"},{"ref":"MK-18064","title":"Geolocated civic problem reports with images","headline":"Geolocated civic problem reports: ~1.3M reports with ~105K images, AI-structured analysis and curator labels","description":"Geolocated reports of physical and digital problems (litter, illegal dumping, infrastructure damage, accessibility barriers, digital issues) from 2023-2026, ~85 GB including images. About 1.3M report records with IDs, submitted descriptions, timestamps and coordinates; ~105K report photos (~77 GB) with an ID manifest; ~1.46M AI-generated analysis rows (titles, descriptions, categories, severity, translations); and curator tags, image checks and annotations. Human-submitted and externally ingested content are distinguishable by source, supporting geographic filtering, issue classification and image/text retrieval.","category":"Geospatial & imagery","sector":"Software & internet","subsector":"Civic issue-reporting platform","dataTypes":["Image","Text","Geospatial","Tabular / structured"],"useCases":["Vision & perception","Retrieval / RAG","Evaluation / benchmarks","World models","Domain fine-tuning"],"origin":"Platform / user-generated","regions":["Global"],"languages":[],"sizeBand":"10-100 GB","count":"~1.3M reports","volume":"84.8 GB; 1,290,899 reports; 104,667 images; 1,458,065 analysis rows","years":"2023–2026","yearsOfHistory":"3","growing":"Yes","systems":["Database"],"contents":["Problem reports (~1.3M) with descriptions","Report photos (~105K)","Latitude, longitude and timestamps","AI-generated titles, categories and severity","Translations","Curator tags and image checks","Source attribution (human vs ingested)"],"formats":["CSV","JSON","SQL","Image files"],"delivery":"","labelled":"Yes","labelTypes":["AI-generated categories and severity","Translations","Curator tags and image checks"],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Civic issue-reporting platform","tools":[["database","Database"]],"span":"3 years","volumeHead":"1.3M reports","sizeTB":0.085,"fromYear":2023,"toYear":2026,"yearsNum":3.0,"tags":["civic issues","geolocated","litter","infrastructure","accessibility","images","severity","crowdsourced"],"shortTitle":"Civic problem reports"},{"ref":"MK-20736","title":"Indian road driving video with GPS and accelerometer (archive + collection to order)","headline":"Indian road dashcam video with GPS and accelerometer: annotated subset (~646K frames) plus archive and fleet collection to order","description":"Dashcam driving data from commercial cab fleets on Indian roads: dense mixed traffic (two-wheelers, auto-rickshaws, trucks, buses, pedestrians, animals) with a high share of night, rain and fog. 1080p/30 fps video synchronised with per-second GPS (position, speed, UTC time) and a ~10 Hz 3-axis accelerometer, G-sensor-triggered event clips, and time-of-day/weather/scene tags. An annotated subset of ~646K frames from ~8.4K clips (~180 hours) carries ~3.7M tracked bounding boxes (12 classes) and ~1.3M segmentation masks (36 classes); a continuous archive sits beyond it. The fleet collects new data to a buyer's spec by city, condition or vehicle type, with LiDAR and radar rigs and human-verified custom labelling available.","category":"Physical-world video","sector":"Automotive & transport","subsector":"Fleet dashcam driving data","dataTypes":["Video","Image","Sensor / telemetry","Geospatial"],"useCases":["Vision & perception","World models","Robotics / embodied AI","Evaluation / benchmarks","Forecasting & time series"],"origin":"Purpose-collected (consented capture)","regions":["South Asia"],"languages":["English","Hindi"],"sizeBand":"","count":"~646K annotated frames (~180 hours)","volume":"Annotated subset: ~646K frames from ~8.4K clips (~180 hours), 3.67M tracked boxes (12 classes), 1.29M segmentation masks (36 classes); continuous archive beyond this; collection scales to order","years":"2026","yearsOfHistory":"","growing":"Yes","systems":["Fleet dashcams","GPS","Accelerometer","Annotation platform"],"contents":["1080p dashcam video (1-minute clips)","Per-second GPS tracks with speed","3-axis accelerometer (~10 Hz)","G-sensor event clips","Tracked bounding boxes (12 classes)","Segmentation masks (36 classes)","Time-of-day, weather and scene tags","Custom collection to spec (LiDAR/radar available)"],"formats":["MP4 (H.265 1080p 30 fps)","JPEG","CSV","TXT","JSON (BDD100K-style)"],"delivery":"Cloud bucket (S3); Cloud bucket (GCS); Physical drives","labelled":"Partly","labelTypes":["Tracked bounding boxes","Segmentation masks","Scene/weather/time-of-day tags","Human-verified custom labels"],"personalData":"Low","deidentification":"Done by owner","rights":"Owner-held","sample":"Available","readiness":"Collected to order","stage":"Inventory ready","scale":"","employees":"","industry":"Fleet dashcam driving data","tools":[],"span":"","volumeHead":"646K annotated frames","sizeTB":null,"fromYear":2026,"toYear":2026,"yearsNum":null,"tags":["driving","dashcam","autonomous vehicles","adas","india","night driving","gps","accelerometer","segmentation","tracking"],"shortTitle":"Indian road driving video"},{"ref":"MK-21745","title":"Consumer AI app (1M+ users): 4 years of private repos with PRs and reviews, plus Figma, Notion, email and agent traces (~6 TB)","headline":"AI chat platform with 1M+ registered users: private web, backend and mobile repos (React Native, SwiftUI) with hundreds of PRs and reviews, plus Figma, Notion, Discord, email, Slack and AI-traces tooling","description":"Engineering and operating record of a consumer AI chat platform: private repositories for web, backend and mobile with full commit, PR and review history, design files, product docs, community channels, email, team chat and AI observability traces.","category":"Code & software","sector":"Software & SaaS","subsector":"Consumer AI app","dataTypes":["Code","Documents","Email","Team chat","Agent trajectories"],"useCases":["Coding models","Agent evaluation"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"1-10 TB","count":"","volume":"~6 TB","years":"~4 years","yearsOfHistory":"","growing":"","systems":["GitHub","Figma","Notion","Discord","Gmail","Slack"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Export 1-2 weeks","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Consumer AI app","tools":[["slack","Slack"],["discord","Discord"],["github","GitHub"],["gmail","Gmail"],["notion","Notion"],["figma","Figma"]],"span":"4 years","volumeHead":"6 TB","sizeTB":6.0,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["codebase","pull requests","code review","ai app"],"shortTitle":"Consumer AI app archive"},{"ref":"MK-22608","title":"Graded coding submissions at scale","headline":"Hundreds of millions of graded coding submissions on thousands of problems with hidden tests, verdicts and edit history","description":"Community data from a large coding platform: a library of thousands of coding questions with hundreds of millions of graded submissions accumulated over 10+ years. Each submission carries the problem statement, constraints and reference solution, hidden test cases with expected outputs and a checker/judge, the submitted code, verdict, runtime, memory and per-test results, edit history between attempts, and difficulty, topic and skill-level tags. Covers 30+ languages, led by Python, Java, SQL dialects, C and C++. Customer data is excluded. Custom collection on long-horizon software tasks through recurring coding events is also possible.","category":"Code & software","sector":"Software & internet","subsector":"Coding challenge platform (community submissions)","dataTypes":["Code","Text","Tabular / structured"],"useCases":["Coding models","Evaluation / benchmarks","RL environments","Preference / RLHF","SFT / instruction tuning"],"origin":"Platform / user-generated","regions":["Global"],"languages":[],"sizeBand":"","count":"Hundreds of millions of graded submissions","volume":"Thousands of questions; hundreds of millions of graded submissions","years":"2026","yearsOfHistory":"10","growing":"Yes","systems":["Custom internal system"],"contents":["Problem statements and constraints","Reference solutions","Hidden test cases with expected outputs","Checkers / judges","Submitted code","Verdicts, runtime and memory","Per-test results","Edit history between attempts","Difficulty, topic and skill-level tags"],"formats":[],"delivery":"","labelled":"Yes","labelTypes":["Verdicts","Per-test results","Runtime and memory","Difficulty ratings","Topic tags","Skill level"],"personalData":"Low","deidentification":"Done by owner","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"Large organisation","employees":"200+","industry":"Coding challenge platform","tools":[],"span":"10 years","volumeHead":"Thousands of questions","sizeTB":null,"fromYear":null,"toYear":2026,"yearsNum":10.0,"tags":["coding","code submissions","test cases","verdicts","programming languages","benchmarks","rl for code","edit history"],"shortTitle":"Graded coding submissions"},{"ref":"MK-22922","title":"Personal and household care contract manufacturer: job folders, formulations and labels","headline":"Contract manufacturer of household and personal care products: ~10K job folders with formulations, labels, POs and inventory","description":"Nine years of records from a small contract manufacturer of household and personal care products. About 10,000 job folders hold invoices, purchase orders, inventory tracking, product formulations and label/artwork designs. Supporting systems include an accounting package (invoices, general ledger, POs), a business email account with supplier correspondence and product artwork, and a custom inventory system (about 3 years) mapping inventory end to end; files are also spread across a cloud drive and chat apps.","category":"Company operations archive","sector":"Manufacturing","subsector":"Contract household & personal care manufacturing","dataTypes":["Documents","Email","Tabular / structured","Image","Chat & messaging"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG","Pre-training"],"origin":"First-party company records","regions":["Southeast Asia"],"languages":[],"sizeBand":"1-10 TB","count":"~10,000 job folders","volume":"~2 TB; ~10,000 job folders","years":"2026","yearsOfHistory":"9","growing":"Yes","systems":["Accounting","Email","Inventory system (custom)","File storage","Team chat"],"contents":["Job folders (~10K)","Product formulations","Label and artwork designs","Invoices","Purchase orders","General ledger","Inventory tracking and end-to-end inventory mapping","Supplier email correspondence"],"formats":["Accounting system exports","Email","Documents and design files"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"Contract household & personal care manufacturing","tools":[["chat","Team chat"],["email","Email"],["accounting","Accounting"],["erp","ERP"],["files","File storage"]],"span":"9 years","volumeHead":"10K job folders","sizeTB":2.0,"fromYear":null,"toYear":2026,"yearsNum":9.0,"tags":["contract manufacturing","personal care","household products","formulations","label artwork","purchase orders","inventory","malaysia"],"shortTitle":"Personal-care manufacturer files"},{"ref":"MK-24137","title":"Egocentric factory manipulation video with tactile gloves","headline":"Egocentric head-camera video of real factory work across ~150 Asian factories, with a tactile-glove (pressure + hand pose) subset","description":"First-person head-camera video of consenting workers doing real manufacturing tasks during normal shifts at ~150 partner factories in India, Vietnam and Indonesia, stated as ~50,000 hours (~170 TB) with collection ongoing. A smaller subset carries synchronised tactile-glove streams (per-sensor pressure and hand pose). Per-recording metadata covers site code, task type, duration, pseudonymous worker ID, camera model and date; annotations include task/step segmentation, success/failure, object labels and hand-visibility flags. Recordings pass automated blur/framing/hand-visibility checks, sampled manual QA, de-duplication and glove time-sync validation; collection can be directed to specific task types.","category":"Physical-world video","sector":"Manufacturing","subsector":"Factory egocentric manipulation capture","dataTypes":["Video","Sensor / telemetry","Tabular / structured"],"useCases":["Robotics / embodied AI","World models","Vision & perception"],"origin":"Purpose-collected (consented capture)","regions":["South Asia","Southeast Asia"],"languages":[],"sizeBand":"100 TB+","count":"~50,000 hours","volume":"~170 TB (~4 GB/hour); ~50,000 hours stated (seller figures to be reconciled)","years":"2026","yearsOfHistory":"","growing":"Yes","systems":["Head-mounted cameras","Tactile data gloves","Cloud storage"],"contents":["Egocentric factory task video","Synchronised tactile-glove pressure streams","Hand-pose streams","Per-recording metadata (site code, task type, worker pseudonym, camera, date)","Task and step segmentation","Task success/failure labels","Object labels","Hand-visibility flags"],"formats":["MP4","CSV","HDF5","JSON"],"delivery":"Cloud bucket (S3); Physical drives","labelled":"Yes","labelTypes":["Task/step segmentation","Success/failure","Object labels","Hand-visibility flags"],"personalData":"Moderate","deidentification":"Done by owner","rights":"Owner-held","sample":"Available","readiness":"2-6 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"Factory egocentric manipulation capture","tools":[["files","File storage"]],"span":"","volumeHead":"50K hours","sizeTB":170.0,"fromYear":2026,"toYear":2026,"yearsNum":null,"tags":["egocentric video","factory","manipulation","tactile glove","hand pose","robotics","vla","imitation learning","world models"],"shortTitle":"Tactile-glove factory video"},{"ref":"MK-24229","title":"Social-bot AI image NFT platform: full source code","headline":"Full codebase for a social-media bot that generates AI images and mints them as NFTs: TypeScript bot server, Solidity contracts with tests, web frontend and git history","description":"Complete source for a small crypto product that takes requests through a X/Twitter bot, generates AI images and mints them as NFTs on an Ethereum layer-2 network. Includes a TypeScript bot server, 7 Solidity smart contracts with Foundry tests, a Next.js frontend and git history with 14 merged pull requests. About 125 files and ~12.5K lines of code across 17 commits, from 2025-2026. Part of the code is MIT-licensed boilerplate and a third-party starter template; the product logic is original. History was squashed in March 2026.","category":"Code & software","sector":"Crypto & web3","subsector":"AI image NFT platform","dataTypes":["Code"],"useCases":["Coding models","Agentic coding"],"origin":"First-party company records","regions":["Not stated"],"languages":["TypeScript","Solidity"],"sizeBand":"<10 GB","count":"~12.5K lines of code","volume":"~125 files; ~12.5K LOC; 17 commits; 14 merged PRs; 7 smart contracts","years":"2025–2026","yearsOfHistory":"","growing":"","systems":["Git"],"contents":["TypeScript bot server","Solidity smart contracts with Foundry tests","Next.js frontend","Git history and merged PRs"],"formats":["Git repository"],"delivery":"","labelled":"No","labelTypes":[],"personalData":"None","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"On request","stage":"Inventory ready","scale":"","employees":"","industry":"AI image NFT platform","tools":[["git","Git"]],"span":"1 year","volumeHead":"12K lines of code","sizeTB":null,"fromYear":2025,"toYear":2026,"yearsNum":1,"tags":["code","solidity","typescript","nft","smart contracts","bot","web3"],"shortTitle":"AI image NFT bot codebase"},{"ref":"MK-24236","title":"Real commercial drone and ground video: ~8,900 clips, ~3.3 TB (construction, industrial, energy, inspection)","headline":"~8,900 real commercial drone and ground videos: construction progress, factories, oil and gas, offshore, heavy lifts, roofing, FPV interiors, roads, bridges and flood damage","description":"Video from real commercial jobs: construction progress, industrial and manufacturing facilities, oil and gas, offshore, heavy lifts, commercial roofing, FPV interior footage, real estate and land, roads, bridges and flood damage. A mix of produced videos and thousands of short source clips, often linked to the project's imagery, mapping and reports.","category":"Physical-world video","sector":"Industrial & engineering","subsector":"Drone / geospatial services","dataTypes":["Video"],"useCases":["Robotics & world models","Computer vision"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"1-10 TB","count":"~8,900 videos","volume":"~3.3 TB; ~8,900 video files","years":"2017–2026","yearsOfHistory":"9","growing":"","systems":[],"contents":[],"formats":["MP4, MOV"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"On request","readiness":"On request","stage":"Inventory ready","scale":"","employees":"","industry":"Drone / geospatial services","tools":[],"span":"9 years","volumeHead":"8.9K videos","sizeTB":3.3,"fromYear":2017,"toYear":2026,"yearsNum":9.0,"tags":["drone video","construction","industrial","inspection"],"shortTitle":"Commercial drone video library"},{"ref":"MK-24247","title":"Career and interview-preparation interaction dataset","headline":"Career and interview-prep interaction data: ~200K profiles with AI feedback, human edits, accept/reject signals and outcomes","description":"First-party data from a career and interview-preparation platform (early 2024-2026): ~200K candidate resumes/profiles linked to target job descriptions, generated interview questions, user answers, AI feedback, before/after resume versions, user corrections to AI-written content, accepted and rejected AI recommendations, and application/interview outcomes where available. About 60-80K candidates have the richest linked candidate-to-job-to-AI-interaction-to-outcome trajectories. Metadata includes extracted skills, job title, seniority, industry and salary expectations. Known gaps: uneven field coverage, near-duplicate resume versions and some self-reported outcomes.","category":"Human preference / evaluation","sector":"Staffing & HR","subsector":"Career and interview-prep platform","dataTypes":["Text","Documents","Tabular / structured"],"useCases":["Preference / RLHF","SFT / instruction tuning","Evaluation / benchmarks","Enterprise workflow agents","Domain fine-tuning"],"origin":"Platform / user-generated","regions":["Global"],"languages":["English"],"sizeBand":"100 GB-1 TB","count":"~200K profiles","volume":"~0.2 TB; ~200K profiles, of which ~60-80K have rich interaction and outcome trajectories","years":"2024–2026","yearsOfHistory":"3","growing":"Yes","systems":["Database"],"contents":["Resumes/candidate profiles (~200K)","Target job descriptions","Generated interview questions","User interview answers","AI feedback on answers","Before/after resume versions","User corrections to AI content","Accepted/rejected AI recommendations","Application and interview outcomes","Skills, title, seniority, industry, salary expectations"],"formats":["Parquet","JSONL","PDF","DOCX"],"delivery":"Cloud bucket (S3)","labelled":"Yes","labelTypes":["Skills, title, seniority, industry metadata","AI feedback","Human edits and corrections","Accept/reject signals","Application/interview outcomes (some self-reported)"],"personalData":"High","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"On request / under NDA","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"Career and interview-prep platform","tools":[["database","Database"]],"span":"3 years","volumeHead":"200K profiles","sizeTB":0.2,"fromYear":2024,"toYear":2026,"yearsNum":3.0,"tags":["career","resumes","interview prep","human feedback","preference data","recruiting","job matching","outcomes"],"shortTitle":"Interview-prep interactions"},{"ref":"MK-25361","title":"Zero-knowledge cryptography tooling: private record","headline":"Zero-knowledge cryptography team: ~1.5M historical email-signing (DKIM) key records, partner chats, issues, meetings and private code","description":"Private operating record (2022-2026) of a zero-knowledge cryptography team building proof tooling for verifying facts from signed web data. Includes a Postgres archive of ~1.5M historical DKIM public keys across ~400K domains with ~320K signature-metadata records (no email bodies; keys back to 2008), ~970 Drive files (specs, grant applications, partner docs, talk decks and recorded talks), ~35K group-chat messages across ~350 partner, audit and hackathon groups, ~35 team chat channels, ~2.3K issues, ~70 recorded meetings (~30 hours) and 20+ private repos (SDK backend, monitoring, infrastructure-as-code, load tests, demos). Public repos are excluded. Offered by an authorised reseller.","category":"Code & software","sector":"Crypto & web3","subsector":"Zero-knowledge cryptography tooling","dataTypes":["Tabular / structured","Code","Documents","Chat & messaging","Email","Video","Text"],"useCases":["Coding models","Domain fine-tuning","Enterprise workflow agents","Retrieval / RAG"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"10-100 GB","count":"~1.5M DKIM key records","volume":"~15-45 GB; ~1.5M signing-key records across ~400K domains; ~35K group-chat messages; ~2K issues; ~70 meetings; 20+ private repos","years":"2022–2026","yearsOfHistory":"4","growing":"","systems":["Database","File storage","Team chat","Issue tracker","Meeting recorder","Email","Git"],"contents":["Historical DKIM public keys by domain/selector (~1.5M)","DKIM signature metadata (~320K)","JWKS snapshots","Specs and grant applications","Recorded talks and meetings","Partner, audit and hackathon group chats (~35K messages)","Engineering chat channels incl. shared partner channels","Issue tracker tickets (~2.3K)","Meeting transcripts (~70 meetings, ~30 h)","Private repos: SDK backend, monitors, infra, load tests, circuits"],"formats":["SQL dump / CSV","DOCX","PPTX","PDF","Video","JSON (chat exports)","MBOX","Git"],"delivery":"Download","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Via authorised intermediary","sample":"","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"Zero-knowledge cryptography tooling","tools":[["postgresql","Postgres"],["chat","Team chat"],["tickets","Tickets"],["git","Git"],["email","Email"],["calls","Calls"],["files","File storage"]],"span":"4 years","volumeHead":"1.5M DKIM key records","sizeTB":0.016,"fromYear":2022,"toYear":2026,"yearsNum":4.0,"tags":["zero-knowledge","zk proofs","dkim","email signatures","cryptography","partner chats","codebase","web3"],"shortTitle":"ZK cryptography tooling"},{"ref":"MK-25753","title":"Former enterprise Microsoft 365 email and Teams archive: ~36.5M items","headline":"~8.3 TB of Exchange mailboxes and Teams communications from a company that formerly had several hundred staff","description":"Historical email and Teams communications of a company that formerly had several hundred employees. The Exchange inventory covers ~760 mailboxes (most inactive), including user, shared, room and scheduling mailboxes, with ~36.5M mailbox items plus ~13.4M recoverable items, mostly under discovery hold. About 8.3 TB in total (~6.1 TB mailbox content, ~1.8 TB recoverable items, ~0.36 TB archives). Likely includes employee, candidate and client correspondence; discovery-hold items may be litigation related. Data owner and signatory are still to be confirmed.","category":"Company operations archive","sector":"Professional services","subsector":"Recruitment platform","dataTypes":["Email","Chat & messaging"],"useCases":["Enterprise workflow agents","Retrieval / RAG","Domain fine-tuning"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"1-10 TB","count":"~36.5M mailbox items","volume":"~8.3 TB; ~760 mailboxes (~180 active, ~580 inactive); ~36.5M mailbox items (~10.5M active, ~26M inactive); ~13.4M recoverable items","years":"","yearsOfHistory":"","growing":"","systems":["Email (Microsoft 365 / Exchange Online)","Team chat (Teams)"],"contents":["User mailboxes","Shared mailboxes","Room and scheduling mailboxes","Recoverable items under discovery hold","Mailbox archives","Teams communications"],"formats":["Microsoft 365 / Exchange Online mailboxes","Teams"],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Inventory ready","scale":"","employees":"","industry":"Recruitment platform","tools":[["email","Outlook"],["chat","Team chat"]],"span":"","volumeHead":"36M mailbox items","sizeTB":8.3,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["email archive","exchange","microsoft 365","teams","enterprise communications","discovery hold"],"shortTitle":"Exchange and Teams mail archive"},{"ref":"MK-26006","title":"ServiceNow and Salesforce consultancy: ~6 TB incl. 1.5 TB of ServiceNow instance data","headline":"Enterprise-platform consultancy: ~6 TB of documents, mailboxes, SharePoint and two ServiceNow instances (~1.5 TB)","description":"IT consultancy implementing ServiceNow and Salesforce for clients. Document storage (~2.8 TB), mailboxes (~1.1 TB plus ~0.3 TB shared), SharePoint (~0.15 TB) and two ServiceNow instances (~1.5 TB). A contact list and candidate CVs would be excluded or de-identified.","category":"Company operations archive","sector":"Software & internet","subsector":"ServiceNow / Salesforce consultancy","dataTypes":["Email","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Computer-use & tool-use agents","Domain fine-tuning"],"origin":"First-party company records","regions":["South Asia","North America"],"languages":[],"sizeBand":"1-10 TB","count":"~6 TB","volume":"~6 TB","years":"","yearsOfHistory":"","growing":"","systems":["Email","File storage","ITSM (ServiceNow)"],"contents":["ServiceNow instance data","Mailboxes","Project documents","SharePoint"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"Planned / on request","rights":"Via authorised intermediary","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Large organisation","employees":"200+","industry":"ServiceNow / Salesforce consultancy","tools":[["tickets","ServiceNow"],["salesforce","Salesforce"],["files","SharePoint"],["email","Email"]],"span":"","volumeHead":"6 TB","sizeTB":6.0,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["servicenow","salesforce","consultancy","itsm","email"],"shortTitle":"ServiceNow consultancy archive"},{"ref":"MK-26120","title":"Occupational exposure and workplace-hazard assessments","headline":"Tens of millions of on-site workplace assessments linking occupations and tasks to measured exposures and hazards, ~20 years","description":"Proprietary structured database of workplace assessments built up over approximately 20 years. It links occupations, work activities, equipment and materials to real-world measurements of occupational exposures, environmental conditions and workplace hazards. Measurements are taken on site by engineers with specialised, certified laboratory equipment at each workplace. Suited to workplace-safety, industrial AI, exposure prediction and risk-assessment models; a database schema can be provided.","category":"Industrial & sensor","sector":"Professional services","subsector":"Occupational health and safety assessment","dataTypes":["Tabular / structured","Sensor / telemetry"],"useCases":["Domain fine-tuning","Forecasting & time series","Retrieval / RAG","Pre-training"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"Tens of millions of assessments","volume":"Tens of millions of assessments","years":"2000–2026","yearsOfHistory":"20","growing":"","systems":["Database","Instruments"],"contents":["Workplace assessments","Occupations and work activities","Equipment and materials used","Measured occupational exposures","Environmental conditions","Workplace hazards"],"formats":["SQL database (exportable)"],"delivery":"Cloud bucket (S3)","labelled":"Partly","labelTypes":[],"personalData":"None","deidentification":"","rights":"Owner-held","sample":"On request / under NDA","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"Occupational health and safety assessment","tools":[["database","Database"]],"span":"20 years","volumeHead":"Tens of millions of assessments","sizeTB":null,"fromYear":2000,"toYear":2026,"yearsNum":20.0,"tags":["occupational safety","workplace hazards","exposure measurements","industrial hygiene","structured database","risk assessment"],"shortTitle":"Workplace hazard assessments"},{"ref":"MK-26389","title":"Live-commerce video archive and graded collectible-card images","headline":"Live-commerce video (~85K hours of collectible-card livestreams) plus ~3.3M graded-card product images","description":"Recorded livestreams from a live-shopping marketplace for collectible trading cards, 2022-2026: ~29.8K streams / ~85.5K hours (median ~2.4 hours each, ~95% English) of hosts presenting, grading and opening cards, running live auctions and breaks, with seller commentary and audience interaction. Plus ~3.28M product photos of graded slabs (grade and certification labels visible), raw cards and sealed product (2021-2026), and product-analytics event streams, marketing campaign/engagement data and web session replays. Payments, payroll, KYC, user PII and addresses and minors' data are excluded; analytics would be de-identified, and seller video needs terms-of-service and seller-consent review before sale.","category":"Video & imagery","sector":"Retail & e-commerce","subsector":"Live-shopping marketplace for collectibles","dataTypes":["Video","Audio / speech","Image","Tabular / structured","Screen recordings / UI traces"],"useCases":["Vision & perception","Speech & audio models","Pre-training","World models","Computer-use & tool-use agents"],"origin":"Platform / user-generated","regions":["North America"],"languages":["English"],"sizeBand":"10-100 TB","count":"~85.5K hours of livestream video","volume":"~40-60 TB (mostly video); ~29.8K streams / ~85.5K hours; ~3.28M images","years":"2021–2026","yearsOfHistory":"5","growing":"","systems":["Cloud storage / CDN","Analytics","Marketing automation","Session replay"],"contents":["Recorded livestream video","Live auctions and card-opening sessions","Seller commentary and audience interaction","Graded-card product photos","Raw card and sealed product photos","Product analytics event streams","Email/push/SMS campaign engagement","Web session replays"],"formats":["MP4 (from HLS)","Images","CSV","API exports"],"delivery":"","labelled":"No","labelTypes":[],"personalData":"Moderate","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"","readiness":"2-6 weeks","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Live-shopping marketplace","tools":[["analytics","Analytics & marketing"],["files","File storage"]],"span":"5 years","volumeHead":"86K hours of livestream video","sizeTB":50.0,"fromYear":2021,"toYear":2026,"yearsNum":5.0,"tags":["livestream","live commerce","video","trading cards","collectibles","product images","auctions","e-commerce"],"shortTitle":"Live-commerce and graded cards"},{"ref":"MK-27528","title":"Project-management SaaS application database (2017-2025)","headline":"Discontinued project-management SaaS database: projects, tasks, comments, activity and templates across many companies (2017-2025)","description":"The full application database of a discontinued project-management SaaS product used by many customer companies, mainly in the US and India. It holds projects with timelines, status and templates; tasks with assignees, dates, completion times and dependencies; comments; per-task activity history; dashboards and saved filtered views. It shows what work companies plan, how long tasks take, what delays them, and how project workflows differ by vertical. The owner will anonymise user and company details before delivery.","category":"Code & software","sector":"Software & internet","subsector":"Project-management SaaS (discontinued)","dataTypes":["Tabular / structured","Text"],"useCases":["Enterprise workflow agents","Computer-use & tool-use agents","RL environments","Forecasting & time series","Pre-training"],"origin":"Platform / user-generated","regions":["North America","South Asia"],"languages":["English"],"sizeBand":"10-100 GB","count":"~2.79M tasks","volume":"~95 GB SQL database: ~5,000 customer companies, ~41.8K projects, ~2.79M tasks","years":"2017–2025","yearsOfHistory":"8","growing":"No","systems":["Database"],"contents":["Projects with timelines and status","Project templates","Tasks with assignees, dates and completion times","Task dependencies","Comments","Task activity history","Dashboards (chart metadata)","Saved filtered views"],"formats":["SQL"],"delivery":"Database dump","labelled":"Yes","labelTypes":[],"personalData":"Moderate","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"Project-management SaaS","tools":[["database","Database"]],"span":"8 years","volumeHead":"2.8M tasks","sizeTB":0.095,"fromYear":2017,"toYear":2025,"yearsNum":8.0,"tags":["project management","saas","tasks","comments","workflows","templates","productivity app","enterprise agents"],"shortTitle":"Project-management SaaS database"},{"ref":"MK-27933","title":"Southeast Asian digital-health company: ~1.5M EMR records linked to ~500K doctor consult chats, plus coaching chats (~8 TB)","headline":"~1.5M electronic medical records (100K+ patients, 2022-2026) linked to ~500K doctor consultation chats, ~250K health-coaching chats with body metrics, plus company documents, SOPs and meeting video","description":"Clinical and operating data from a venture-backed digital-health company: electronic medical records for 100K+ patients, doctor consultation chats linked to those records, health-coaching chats with body metrics, CRM, and the company's documents, SOPs, meeting recordings and email. Mostly in Bahasa Indonesia. Licensable only as a de-identified set.","category":"Healthcare & life sciences","sector":"Healthcare & life sciences","subsector":"Telemedicine","dataTypes":["Text","Chat & messaging","Tabular / structured","Video","Email","Documents"],"useCases":["Medical models","Multilingual","Conversational agents"],"origin":"First-party company records","regions":["Southeast Asia"],"languages":[],"sizeBand":"1-10 TB","count":"~1.5M EMR records","volume":"~8 TB; ~1.5M EMR records; ~500K consult chats; ~250K coaching chats; 690 GB meeting video; 257 GB email","years":"2022–2026","yearsOfHistory":"4","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"Required before release","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Mid-size company","employees":"51–200","industry":"Telemedicine","tools":[],"span":"4 years","volumeHead":"1.5M EMR records","sizeTB":8.0,"fromYear":2022,"toYear":2026,"yearsNum":4.0,"tags":["health","emr","doctor chats","bahasa indonesia","telemedicine"],"shortTitle":"Digital-health records and chats"},{"ref":"MK-27969","title":"Consumer tech startup: 8 years of Slack, Jira, Notion, GitHub, Drive and meeting recordings","headline":"Full cross-linked operating record of a consumer app and game company: ~4M chat messages, 15k tickets with worklogs, 53k commits linked to tickets, 30k wiki pages","description":"The complete internal record of a venture-backed consumer technology company that ran a mobile game and a life-organiser app, covering 2018 to 2026 with code history from 2014. It spans team chat (about 4M messages across 800+ public channels), an issue tracker (15k issues and 31k worklogs), wiki and knowledge-base pages (about 31k), 193 code repositories with 53k commits, a 836 GB shared drive (product specs, roadmaps, OKRs, financial models, board decks, design and data pipelines), about 82 hours of internal meeting recordings and customer support history. Issue keys in commit messages link tickets to the code that resolved them, giving ticket-to-commit software engineering trajectories, and docs, threads and tickets reference each other throughout. A pseudonymised knowledge-base slice (about 15.9k pages and 1.2k structured tables) is already prepared. Employee names, messages and meeting recordings need pseudonymisation and consent, and code needs a secrets scrub before delivery. Email is available on request.","category":"Company operations archive","sector":"Software & internet","subsector":"Consumer mobile app and game developer","dataTypes":["Chat & messaging","Tickets","Source code","Documents","Spreadsheets","Video","Audio / speech"],"useCases":["Agent environments","Long-horizon tasks","Retrieval / RAG","Software engineering","Evals","Domain fine-tuning"],"origin":"Company internal records","regions":["North America","Europe"],"languages":["English"],"sizeBand":"100 GB-1 TB","count":"~4M chat messages","volume":"~1 TB; ~4M chat messages (800+ public channels); 15k issues and 31k worklogs; ~31k wiki pages; 193 repos and 53k commits; 836 GB shared drive; ~82 h of meeting recordings","years":"2018–2026","yearsOfHistory":"8","growing":"","systems":["Slack","Jira","Notion","Confluence","GitHub","Google Drive","Intercom","Dropbox","Gmail"],"contents":["Decision threads, launches and incidents","Tickets with worklogs","Ticket-to-commit links","PRDs, specs, OKRs and roadmaps","Source code and commit history","Financial models and board decks","Internal meeting recordings","Game economy and LiveOps playbooks"],"formats":["JSON","CSV","Markdown","Git","Office documents","Video"],"delivery":"Per-system exports (JSON, CSV, Markdown, Git), pseudonymised and secrets-scrubbed, with SHA-256 manifest","labelled":"Partly","labelTypes":["Ticket-to-commit links","Time spent per ticket","Structured epics, user stories and QA tables"],"personalData":"Moderate","deidentification":"Pseudonymisation and secret scrubbing planned; knowledge-base slice already filtered","rights":"To be confirmed","sample":"Available on request","readiness":"On request","stage":"Inventory ready","scale":"","employees":"","industry":"Consumer mobile app and game developer","tools":[["slack","Slack"],["github","GitHub"],["jira","Jira"],["intercom","Intercom"],["gmail","Gmail"],["notion","Notion"],["confluence","Confluence"],["googledrive","Google Drive"],["dropbox","Dropbox"]],"span":"8 years","volumeHead":"4M chat messages","sizeTB":1.0,"fromYear":2018,"toYear":2026,"yearsNum":8.0,"tags":["company in a box","startup","slack","jira","github","ticket to commit","swe trajectories","meeting recordings","consumer app","game studio"],"shortTitle":"Consumer-tech startup operating record"},{"ref":"MK-28602","title":"Wine producer: 10 years of email, documents, accounting, e-commerce and CRM","headline":"California wine producer (DTC, wholesale, retail, hospitality): 10 years of email, ~125 GB documents, ~7K accounting transactions, ~9K orders, ~23K CRM records","description":"Small wine producer selling direct-to-consumer, wholesale, retail and through hospitality. Email (~46 GB, 5 inboxes), documents (~125 GB), accounting (~7.2K transactions), e-commerce (~8.8K orders), email-marketing CRM (~23K customer records) and AI-assistant workspace files.","category":"Company operations archive","sector":"Hospitality & food","subsector":"Winery","dataTypes":["Email","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"100 GB-1 TB","count":"","volume":"~170 GB (email covers ~10 years)","years":"2016–2026","yearsOfHistory":"10","growing":"","systems":["Email","File storage","Accounting","E-commerce","CRM"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"Winery","tools":[["email","Email"],["crm","CRM"],["accounting","Accounting"],["shop","Commerce"],["files","File storage"]],"span":"10 years","volumeHead":"170 GB","sizeTB":0.17,"fromYear":2016,"toYear":2026,"yearsNum":10.0,"tags":["wine","dtc","e-commerce","accounting"],"shortTitle":"Winery operating archive"},{"ref":"MK-28676","title":"Restaurant guest-service conversations","headline":"Restaurant guest-service conversations from an AI concierge: ~131K multi-channel chats with reservations and AI tool calls","description":"Real conversations between ~95 restaurant venues and their guests handled by an AI customer-service platform across WhatsApp, Instagram, Messenger, email, SMS and phone: 130,840 conversations and 630,576 messages (WhatsApp ~95K, Instagram ~14K, email ~5K, SMS ~1.7K), 32,736 AI tool calls (e.g. availability checks, bookings), 15,021 reservations and human-takeover replies. Also restaurant knowledge bases, review replies and call logs. Delivered as anonymised JSONL, one complete conversation per record, with personal details replaced by tokens; a regular feed of new conversations can be set up.","category":"Speech & conversation","sector":"Hospitality & food","subsector":"Restaurant AI customer service / reservations","dataTypes":["Chat & messaging","Text","Agent trajectories","Tabular / structured"],"useCases":["SFT / instruction tuning","Enterprise workflow agents","Computer-use & tool-use agents","Domain fine-tuning","Evaluation / benchmarks"],"origin":"Platform / user-generated","regions":["Europe"],"languages":[],"sizeBand":"<10 GB","count":"130,840 conversations","volume":"130,840 conversations; 630,576 messages; 32,736 AI tool calls; 15,021 reservations; ~95 venues","years":"2026","yearsOfHistory":"2","growing":"Yes","systems":["Database","Messaging channels","Telephony"],"contents":["Multi-channel guest conversations","AI tool-call logs","Reservations","Human-takeover replies","Restaurant knowledge bases","Review replies","Call logs"],"formats":["JSONL"],"delivery":"","labelled":"Partly","labelTypes":["AI actions","Human takeovers"],"personalData":"Low","deidentification":"Done by owner","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"Restaurant AI customer service / reservations","tools":[["whatsapp","WhatsApp"],["calls","Calls"],["database","Database"]],"span":"2 years","volumeHead":"131K conversations","sizeTB":0.002,"fromYear":null,"toYear":2026,"yearsNum":2.0,"tags":["restaurants","customer service","whatsapp","conversational ai","tool calls","reservations","chat logs","hospitality"],"shortTitle":"Restaurant guest conversations"},{"ref":"MK-30344","title":"US commercial drone and geospatial firm: ~5 TB, 2017-2026 (imagery, thermal, mapping, CAD, email, accounting)","headline":"Nearly a decade of a US drone/geospatial services firm's real jobs: aerial and close-range imagery and video, thermal/IR inspections, orthomosaics, CAD and point clouds, survey data, reports, ~144 GB email and accounting","description":"Complete operating archive of a US commercial drone and geospatial services company: aerial and close-range RGB imagery and video, thermal/IR inspections (schools, compressor stations, coastal industrial, roofing, solar), mapping and orthomosaics, CAD/engineering files, point clouds and 3D models, survey data, inspection reports, project documents, ~144 GB of email, Microsoft 365 files and accounting records. Client and project identities withheld.","category":"Company operations archive","sector":"Industrial & engineering","subsector":"Drone / geospatial services","dataTypes":["Images","Video","CAD / 3D","Geospatial","Email","Documents","Tabular / structured"],"useCases":["Robotics & world models","Computer vision","Enterprise workflow agents"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"1-10 TB","count":"","volume":"~5 TB consolidated; thermal ~141 GB; mapping ~167 GB; construction progress ~307 GB; offshore energy ~79 GB; oil and gas ~52 GB; email ~144 GB","years":"2017–2026","yearsOfHistory":"9","growing":"","systems":["Pix4D","DroneDeploy","CAD/BIM","ReCap","GIS","Outlook / 365","QuickBooks"],"contents":[],"formats":["JPG, DNG/RAW, MP4/MOV, GeoTIFF, SHP/KML, DWG/DXF/RVT/SKP, RCS/LAS/OBJ, PST, Office, PDF"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Medium","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"Anonymised samples on request","readiness":"On request","stage":"Inventory ready","scale":"","employees":"","industry":"Drone / geospatial services","tools":[["email","Outlook"],["quickbooks","QuickBooks"],["design","Design files"]],"span":"9 years","volumeHead":"5 TB consolidated","sizeTB":5.0,"fromYear":2017,"toYear":2026,"yearsNum":9.0,"tags":["drone","geospatial","thermal","point cloud","cad","construction","inspection"],"shortTitle":"Drone and geospatial firm archive"},{"ref":"MK-30503","title":"US pet-care services business: full operating archive","headline":"US pet-care services business: ~5 years of CRM, ~100K call recordings, email, telephony, scheduling and cloud data","description":"Exports from a US pet-care services business about five years old (~420 GB in total). A CRM export (~230 GB) holds ~6.4M records, ~596K messages, ~74K call recordings and ~77K files; a workspace takeout (~95 GB) covers drive, email and groups for 231 accounts; a franchise-system CRM org adds records, attachments and a database copy (~22 GB); telephony adds ~28K call recordings plus call and text logs (~15 GB). Also e-signature documents (~8.8K), cloud databases, logs, code and settings, workforce timesheets, schedules and employee data, one year of team chat (~900 conversations, ~20K files) and accounting.","category":"Company operations archive","sector":"Consumer services","subsector":"Pet daycare and care services","dataTypes":["Audio / speech","Email","Chat & messaging","Documents","Tabular / structured","Code"],"useCases":["Speech & audio models","Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG","Pre-training"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"100 GB-1 TB","count":"~100K call recordings","volume":"CRM 230 GB (6.4M records, 596K messages, 74K call recordings); email/docs 95 GB; franchise CRM 22 GB; telephony 15 GB (28K recordings); documents 5 GB (8,791); other systems ~50 GB","years":"2026","yearsOfHistory":"5","growing":"Yes","systems":["CRM","Email","File storage","Telephony","E-signature","Database","Cloud logs","Workforce scheduling","Team chat","Accounting"],"contents":["CRM records (~6.4M) and messages (~596K)","CRM call recordings (~74K)","Telephony call recordings (~28K) and call/text logs","Email, drive and groups for 231 accounts","Franchise-system CRM records and attachments","E-signature documents (~8.8K)","Cloud databases, logs and code","Timesheets, schedules and employee data","Team chat (1 year)","Accounting records"],"formats":["Platform exports","PDF","Audio recordings"],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Inventory ready","scale":"","employees":"","industry":"Pet daycare and care services","tools":[["chat","Team chat"],["email","Email"],["crm","CRM"],["calls","Calls"],["database","Database"],["accounting","Accounting"],["calendar","Calendar"],["sign","E-signature"],["files","File storage"]],"span":"5 years","volumeHead":"100K call recordings","sizeTB":0.42,"fromYear":null,"toYear":2026,"yearsNum":5.0,"tags":["pet care","call recordings","crm","customer service","telephony","scheduling","franchise","services business"],"shortTitle":"Pet-care business archive"},{"ref":"MK-30902","title":"AI-solutions firm: email, documents and SharePoint (~260 GB, since 2025)","headline":"AI-solutions services firm: ~260 GB of email, documents and SharePoint from 2025","description":"AI-solutions services company: email (~24 GB), OneDrive (~200 GB) and SharePoint (~38 GB). About one year of history.","category":"Company operations archive","sector":"AI & data","subsector":"AI services","dataTypes":["Email","Documents"],"useCases":["Enterprise workflow agents"],"origin":"First-party company records","regions":["South Asia"],"languages":[],"sizeBand":"100 GB-1 TB","count":"","volume":"~260 GB","years":"2025–2026","yearsOfHistory":"1","growing":"","systems":["Email","File storage"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Via authorised intermediary","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"AI services","tools":[["files","SharePoint"],["files","OneDrive"],["email","Email"]],"span":"1 year","volumeHead":"260 GB","sizeTB":0.26,"fromYear":2025,"toYear":2026,"yearsNum":1.0,"tags":["ai services","email","documents"],"shortTitle":"AI-solutions firm records"},{"ref":"MK-30974","title":"European B2B performance-marketing agency: 8 years of Slack (~711K msgs), PM tasks (~270K), CRM, email, ~3.5K meeting transcripts","headline":"Remote B2B marketing agency: Slack ~711K messages, ~270K project tasks with ~111K comments, CRM, 106 GB email, ~3.5K meeting transcripts, ~490 decks","description":"Full operating record of a European B2B performance-marketing agency working for SaaS clients (paid acquisition, demand generation, creative, landing pages): shared drive, team chat, email, sales CRM, project management with task comments, meeting transcripts, video reviews, surveys and pitch decks. ~99% English. Client-NDA material, client ad accounts, HR, payroll and raw contact data are excluded.","category":"Company operations archive","sector":"Professional services","subsector":"Marketing agency","dataTypes":["Team chat","Email","Documents","Tickets","Meeting transcripts","Tabular / structured"],"useCases":["Enterprise workflow agents","Long-horizon office tasks"],"origin":"First-party company records","regions":["Europe"],"languages":[],"sizeBand":"1-10 TB","count":"~711K messages; ~270K tasks","volume":"Drive 1.71 TB gross; Slack ~711K msgs; email 106 GB; CRM ~52.5K companies, ~1.1K deals; PM ~269.6K tasks; ~3.5K transcripts","years":"~6-9 years by system","yearsOfHistory":"","growing":"","systems":["Google Drive","Slack","Google Workspace","CRM","Project management"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"Owner-side redaction","rights":"Owner-held","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Marketing agency","tools":[["slack","Slack"],["googledrive","Google Drive"],["tickets","Tickets"],["crm","CRM"]],"span":"9 years","volumeHead":"711K messages","sizeTB":1.8,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["agency","marketing","slack","project management","transcripts"],"shortTitle":"B2B marketing agency archive"},{"ref":"MK-32956","title":"Rollup-as-a-service blockchain platform: monorepo and operations","headline":"Rollup-as-a-service blockchain platform: private monorepo (~3,500 PRs) with linked tickets, email, docs and chat","description":"Four years of private engineering and operating history from a venture-backed rollup-as-a-service company that launches and operates Ethereum-settled chains. Core asset is a private monorepo (~3,500 PRs) for chain deployment and orchestration across major rollup stacks, cross-chain messaging, token infrastructure and DevOps, with issue-tracker tickets linked to PRs (est. 2K-10K) and the surrounding team discussion; ~270 further private repos are mostly small partner experiments. Also company email (~20-40 inboxes), Drive specs, decks, contracts, customer onboarding docs and finance sheets, team chat (est. 50K-500K messages) and partner/customer chain chats. Sizes are estimates, not yet measured. Offered by an authorised reseller.","category":"Code & software","sector":"Crypto & web3","subsector":"Rollup-as-a-service blockchain infrastructure","dataTypes":["Code","Documents","Email","Chat & messaging"],"useCases":["Coding models","Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"","count":"~3,500 PRs","volume":"Est. 40-700 GB; ~3,500 PRs in main repo; est. 2K-10K tickets; est. 50K-500K chat messages","years":"2022–2026","yearsOfHistory":"4","growing":"","systems":["Git","Issue tracker","Email","File storage","Team chat"],"contents":["Private monorepo for chain deployment/orchestration (~3,500 PRs)","Cross-chain messaging and token infrastructure code","DevOps configuration","Issue-tracker tickets linked to PRs","Partner-experiment repos (~270)","Company email","Specs, decks and customer onboarding docs","Customer contracts and finance sheets","Engineering team chat","Partner/customer chain chats"],"formats":["Git mirror","CSV","MBOX","DOCX","XLSX","PDF","JSON (chat exports)"],"delivery":"Download","labelled":"Partly","labelTypes":["Tickets linked to PRs"],"personalData":"Moderate","deidentification":"","rights":"Via authorised intermediary","sample":"","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Rollup-as-a-service blockchain infrastructure","tools":[["chat","Team chat"],["tickets","Tickets"],["git","Git"],["email","Email"],["files","File storage"]],"span":"4 years","volumeHead":"3.5K PRs","sizeTB":null,"fromYear":2022,"toYear":2026,"yearsNum":4.0,"tags":["blockchain","rollups","ethereum","devops","monorepo","pull requests","infrastructure","web3"],"shortTitle":"Rollup blockchain platform"},{"ref":"MK-33458","title":"IT and BPO services firm: email, files, accounting and Teams since 2018 (~0.5 TB)","headline":"IT and business-process outsourcing firm: ~0.5 TB of email, files, accounting and Teams since 2018","description":"IT and BPO services company: email (~365 GB), OneDrive and shared drives (~86 GB), accounting (~38 GB), Teams and HR systems.","category":"Company operations archive","sector":"Professional services","subsector":"IT and BPO services","dataTypes":["Email","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"100 GB-1 TB","count":"","volume":"~0.5 TB","years":"2018–2026","yearsOfHistory":"8","growing":"","systems":["Email","File storage","Accounting","Team chat"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"Planned / on request","rights":"Via authorised intermediary","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Mid-size company","employees":"51–200","industry":"IT and BPO services","tools":[["files","OneDrive"],["chat","Team chat"],["email","Email"],["accounting","Accounting"]],"span":"8 years","volumeHead":"0.5 TB","sizeTB":0.5,"fromYear":2018,"toYear":2026,"yearsNum":8.0,"tags":["bpo","it services","email"],"shortTitle":"IT and BPO services records"},{"ref":"MK-33823","title":"Indian home-construction advisory platform: 6-year archive","headline":"Indian home-construction advisory platform: 6 years of chat, CRM deals with outcomes, calls, project media and code","description":"Six years of operating data from a home-construction advisory platform for homeowners in India (consumer app, paid advisory such as quote analysis, design moodboards and plan reviews, and a contractor marketplace). Includes ~800K team-chat messages, email and Drive documents (~5K PDFs of quotes, reports and plans, plus docs and sheets), ~140K CRM contacts and ~5.6K advisory deals with stages and outcomes (~800 won), ~700 hours of recorded advisory/webinar calls and ~11K recorded phone calls, business WhatsApp conversations, a production database and ~3B product events, ~38 TB of project photos and site-progress video, ~0.5M lines of production code, project boards and design files. Recordings are mostly in a regional Indian language and English. The owner exports de-identified data; no live system access.","category":"Company operations archive","sector":"Construction & real estate","subsector":"Home-construction advisory platform and contractor marketplace","dataTypes":["Chat & messaging","Email","Documents","Audio / speech","Video","Image","Tabular / structured","Code"],"useCases":["Enterprise workflow agents","Speech & audio models","Vision & perception","Domain fine-tuning","Coding models","Retrieval / RAG"],"origin":"First-party company records","regions":["South Asia"],"languages":["English","Regional Indian language"],"sizeBand":"10-100 TB","count":"~3B product events","volume":"~42 TB (mostly project media) plus ~4 TB event archive; ~800K chat messages; ~5K PDFs; ~140K CRM contacts; ~5.6K advisory deals (~800 won); ~700 hours of recorded sessions + ~11K phone calls; ~3B product events; ~0.5M LOC","years":"2020–2026","yearsOfHistory":"6","growing":"Yes","systems":["Email","File storage","Team chat","CRM","Telephony","WhatsApp business","Database","Analytics","Cloud storage","Git","Project management","Design files","Accounting"],"contents":["Team chat (~800K messages)","Contractor quotes and quote-comparison reports","BOQs, floor plans and drawings","Design moodboards","CRM contacts and advisory deals with outcomes","Recorded advisory/webinar sessions (~700 h)","Recorded sales/advisory phone calls (~11K)","WhatsApp business conversations","Production app database","Product events (~3B)","Project photos and site-progress video (~38 TB)","Production code (~0.5M LOC)","Project boards and design files"],"formats":["MBOX","Drive/Takeout","JSON (chat)","CSV/SQL","Audio","Media objects","Git"],"delivery":"Download","labelled":"Partly","labelTypes":["Deal stages and outcomes"],"personalData":"High","deidentification":"Done by owner","rights":"Owner-held","sample":"","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Home-construction advisory platform and contractor marketplace","tools":[["whatsapp","WhatsApp"],["tickets","Tickets"],["git","Git"],["email","Email"],["crm","CRM"],["calls","Calls"],["database","Database"],["accounting","Accounting"],["design","Design files"],["hr","HR & payroll"],["analytics","Analytics & marketing"],["files","File storage"]],"span":"6 years","volumeHead":"3B product events","sizeTB":46.0,"fromYear":2020,"toYear":2026,"yearsNum":6.0,"tags":["construction","home building","advisory","india","call recordings","whatsapp","site photos","crm"],"shortTitle":"Home-construction advisor records"},{"ref":"MK-34763","title":"Enterprise marketing-stack workflow environment","headline":"Enterprise marketing-stack RL environment: 1,547 cross-system repair tasks, 426 tools, 100% programmatic verifiers","description":"Simulated connected enterprise marketing stack (ad platforms, mobile attribution, customer data platform, segmentation, journeys, experimentation, email delivery, reporting, remote config and DNS) rebuilt as self-contained emulators. Tool-using agents investigate and repair cross-system workflows across 1,547 tasks using 426 typed JSON tools; deterministic verifiers grade application state, logs and downstream effects with partial credit (no LLM judge). Long-horizon episodes run to dozens of tool calls; early results suggest pass@5 of 30-40%. Delivered as versioned containerised task packages with initial state, verifiers, reference solutions, audit packs and documentation, CPU-only and run on buyer infrastructure.","category":"RL environment / benchmark","sector":"AI & data","subsector":"RL environment: marketing technology workflows","dataTypes":["RL environment / simulator","Agent trajectories","Tabular / structured"],"useCases":["RL environments","Evaluation / benchmarks","Enterprise workflow agents","Computer-use & tool-use agents"],"origin":"Expert-created","regions":["Not stated"],"languages":["English"],"sizeBand":"","count":"1,547 tasks","volume":"1,547 tasks; 426 tools","years":"2026","yearsOfHistory":"","growing":"","systems":["Docker containers","Database"],"contents":["1,547 workflow-repair tasks","426 typed JSON tools","Emulated ad, attribution, CDP and messaging systems","Deterministic programmatic verifiers","Reference solutions","Audit packs","Seeded initial states","Setup and agent-facing documentation"],"formats":["Harbor task packages","Linux containers","SQLite","JSON"],"delivery":"Buyer-hosted container","labelled":"Yes","labelTypes":["Reference solutions","Audit packs"],"personalData":"None","deidentification":"Not needed","rights":"Owner-held","sample":"Available","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"RL environment: marketing technology workflows","tools":[["docker","Docker"],["database","Database"]],"span":"","volumeHead":"1.5K tasks","sizeTB":null,"fromYear":2026,"toYear":2026,"yearsNum":null,"tags":["rl environment","marketing technology","enterprise workflows","tool use","agents","verifiable rewards","long-horizon"],"shortTitle":"Marketing-stack environment"},{"ref":"MK-36119","title":"Healthcare-administration RL environment: 20 simulated hospital and health-plan apps, 778 workflows","headline":"Healthcare-admin RL environment: 20 simulated hospital and health-plan apps, 87 tools, 778 workflows in 179 task categories with state-based verifiers","description":"Simulated hospital and health-plan software (20 apps, 87 tools) with 778 workflows across 179 task categories, ~1.3K handbook documents and state-based verifiers. Frontier models pass 15-37% of tasks first time. Runs in Docker or managed cloud.","category":"RL environment / benchmark","sector":"Healthcare & life sciences","subsector":"Healthcare administration","dataTypes":["RL environment / simulator"],"useCases":["RL environments","Enterprise workflow agents","Evaluation / benchmarks"],"origin":"Generated / simulated","regions":["Not stated"],"languages":[],"sizeBand":"","count":"778 workflows","volume":"778 workflows; 87 tools; ~1.3K handbook docs","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"Yes","labelTypes":[],"personalData":"None","deidentification":"Not needed","rights":"Owner-held","sample":"","readiness":"On request","stage":"Inventory ready","scale":"","employees":"","industry":"Healthcare administration","tools":[["docker","Docker"]],"span":"","volumeHead":"778 workflows","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["rl environment","healthcare","prior auth","workflows","verifiers"],"shortTitle":"Healthcare admin environment"},{"ref":"MK-37785","title":"Hard verifiable maths RL tasks and terminal-based science tasks (1K each)","headline":"1,000 verifiable maths RL tasks in three difficulty tiers (hardest tier: <30% frontier pass rate) plus 1,000 terminal-based science tasks","description":"1,000 maths tasks, each with instructions, a verifier and a reference solution, in three tiers (a leading frontier model passes ~95% of Medium, ~65% of Hard and under 30% of Ultra-Hard), plus 1,000 terminal-based science tasks in a standard agent-benchmark format.","category":"RL environment / benchmark","sector":"AI & data","subsector":"Maths and science RL tasks","dataTypes":["RL environment / simulator","Text"],"useCases":["RL environments","Evaluation / benchmarks"],"origin":"Expert-created","regions":["Europe"],"languages":[],"sizeBand":"","count":"2K tasks","volume":"1K maths tasks; 1K science tasks","years":"2026","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"Yes","labelTypes":[],"personalData":"None","deidentification":"Not needed","rights":"Owner-held","sample":"Available","readiness":"On request","stage":"Inventory ready","scale":"","employees":"","industry":"Maths and science RL tasks","tools":[],"span":"","volumeHead":"2K tasks","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["math","rl","verifiers","science","terminal tasks"],"shortTitle":"Verifiable maths and science tasks"},{"ref":"MK-40369","title":"Global public-records legal corpus","headline":"Global public-records legal corpus: ~490M source records and ~160M files (~75 TB) across US, EU, international law and 84 countries","description":"Legal, regulatory and administrative material gathered from public sources: a US collection (~133M records: federal rulemaking comments, regulatory code sections, court opinions, municipal codes, state statutes and legislation, FOIA releases, veterans' appeals decisions, patent and trademark records), an EU institutional collection (~359M records: legal texts, procurement notices, financial reference entries, regulatory and environmental measurements), ~1.5M public-international-law records (tribunals, treaties, UN documents) and national collections for 84 countries with >10K rows each. Also securities filings, environmental, immigration, sanctions, tax/nonprofit and spectrum records. ~160M stored files (~68 TB; mostly PDF, plus HTML, RTF, text, XML, Word, images and some audio) plus a ~6.6 TB structured database of 797 tables with source IDs, titles, dates, source links and inline text. Can be parsed or converted for pre-training or RL environments.","category":"Text & documents","sector":"Legal","subsector":"Public legal and regulatory records aggregation","dataTypes":["Text","Documents","Tabular / structured"],"useCases":["Pre-training","Retrieval / RAG","Domain fine-tuning","RL environments","Evaluation / benchmarks"],"origin":"Public / scraped sources","regions":["Global"],"languages":[],"sizeBand":"10-100 TB","count":"~490M source records","volume":"133.2M US records; 358.9M EU records; 1.5M international-law records; 160M stored files (~68 TB) plus ~6.6 TB structured database","years":"2026","yearsOfHistory":"","growing":"Yes","systems":["Database","File storage"],"contents":["Federal rulemaking public comments","Court opinions and decisions","Statutes, codes and municipal ordinances","EU legal texts and procurement notices","International treaties and tribunal records","Patent and trademark records","Securities filings and financial facts","Environmental regulation and measurements","Immigration and sanctions records","National legal collections for 84 countries","Structured fields: source IDs, titles, dates, links"],"formats":["PDF","HTML","RTF","TXT","XML","DOCX","JSON","Database tables"],"delivery":"","labelled":"Partly","labelTypes":["Structured fields per dataset"],"personalData":"Moderate","deidentification":"","rights":"Public sources (per-source licences)","sample":"","readiness":"On request","stage":"Inventory ready","scale":"","employees":"","industry":"Public legal and regulatory records aggregation","tools":[["database","Database"],["files","File storage"]],"span":"","volumeHead":"490M source records","sizeTB":74.7,"fromYear":null,"toYear":2026,"yearsNum":null,"tags":["legal","regulatory","court records","public records","multilingual","patents","eu law","pre-training corpus","rag"],"shortTitle":"Public-records legal corpus"},{"ref":"MK-41740","title":"Regional pharmacy retail chain and wholesaler: 10 years of sales lines, baskets, prices and stock","headline":"~65M retail and wholesale sales lines with receipt-level baskets, full SKU price history and per-store stock from a regional pharmacy chain and wholesaler","description":"Ten years of sales and inventory data, about 2 TB, from a regional pharmacy retail chain and wholesaler with over 1M customers and more than 20 years of trading. Includes ~65M sales lines (about 55% wholesale and 45% retail, ~750K per month), receipt-level retail baskets with timestamps per store, ~15K SKUs with full price history, list versus sold price, and historical stock per SKU per store. Pharmacy purchases can reveal health conditions, so the absence of customer IDs or loyalty links is to be confirmed. Ownership is stated by the technical lead rather than the owner.","category":"Consumer behaviour","sector":"Retail & e-commerce","subsector":"Pharmacy retail and wholesale","dataTypes":["Tabular / structured"],"useCases":["Forecasting & time series","Pricing & demand models","Recommendation"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"1-10 TB","count":"~65M sales lines","volume":"~2 TB; ~65M sales lines (~55% wholesale, ~45% retail, ~750K per month); ~15K SKUs with price history; per-store stock history","years":"","yearsOfHistory":"10","growing":"","systems":[],"contents":["Retail and wholesale sales lines","Receipt-level baskets with timestamps per store","SKU catalogue with full price history","List versus sold price","Historical stock per SKU per store"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Mid-size company","employees":"51–200","industry":"Pharmacy retail and wholesale","tools":[],"span":"10 years","volumeHead":"65M sales lines","sizeTB":2.0,"fromYear":null,"toYear":null,"yearsNum":10.0,"tags":["pharmacy","retail","wholesale","transactions","market baskets","pricing","inventory","demand forecasting"],"shortTitle":"Pharmacy chain sales and pricing"},{"ref":"MK-41852","title":"Publishing and media group: 20-year internal archive","headline":"Large European book and media publishing group: ~20 years of internal email, CRM, ledger and video (~1.5 PB stated)","description":"Internal records of a large European publishing group selling books, media and entertainment. Contents described include internal correspondence between colleagues and back-and-forth with authors, videos, CRM records (invoices and payments, ~20 years) and the general ledger held in file storage, across ERP, CRM, servers and email. The submitter states roughly 1.5 PB in total and thousands of employee mailboxes; per-system volumes have not been itemised.","category":"Company operations archive","sector":"Marketing, media & publishing","subsector":"Book and media publishing group","dataTypes":["Email","Documents","Video","Tabular / structured"],"useCases":["Pre-training","Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["Europe"],"languages":[],"sizeBand":"100 TB+","count":"~1.5 PB (stated)","volume":"~1,500 TB (stated); ~2,300 mailboxes","years":"2026","yearsOfHistory":"20","growing":"Yes","systems":["ERP","CRM","Email","File storage","Servers"],"contents":["Employee email correspondence","Editor-author correspondence","Video files","CRM invoices and payments","General ledger"],"formats":["Email","CRM exports","Ledger files","Video"],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"2-6 weeks","stage":"Inventory ready","scale":"Small team","employees":"200+","industry":"Book and media publishing group","tools":[["email","Email"],["crm","CRM"],["erp","ERP"],["files","File storage"]],"span":"20 years","volumeHead":"1.5 PB","sizeTB":1500.0,"fromYear":null,"toYear":2026,"yearsNum":20.0,"tags":["publishing","books","media","author correspondence","email archive","video","general ledger","europe"],"shortTitle":"Publishing group archive"},{"ref":"MK-44836","title":"Specialty chemicals R&D and toll manufacturing: 20-year record","headline":"Specialty chemicals R&D and toll manufacturer: ~20 years linking orders, HPLC/GC/IR runs, COAs and bilingual sourcing email","description":"Records of a 30-year-old specialty chemicals company doing custom synthesis, process development, scale-up and contract manufacturing (~4 TB). Orders and commercial records (~2,000-10,000 orders since ~2016) with linked POs, quotes, invoices, packing lists, freight and customs paperwork across a few hundred customer and supplier accounts, many involving US-China sourcing; in-house HPLC, GC and IR raw instrument files and reports (thousands of runs), outsourced NMR reports and certificates of analysis for most lots; thousands of SDSs, specifications and hazmat/import-export compliance documents; ~20 years of email across several mailboxes, much of it bilingual English/Chinese; and bookkeeping for the full period. An inventory system adds chemical movements, SOPs and an audit log of user actions, so a lot can be traced from RFQ to supplier to chromatogram to COA to shipment.","category":"Company operations archive","sector":"Manufacturing","subsector":"Specialty chemicals custom synthesis & toll manufacturing","dataTypes":["Documents","Email","Tabular / structured","Sensor / telemetry"],"useCases":["Domain fine-tuning","Enterprise workflow agents","Pre-training","Retrieval / RAG","Evaluation / benchmarks"],"origin":"First-party company records","regions":["North America"],"languages":["English","Chinese"],"sizeBand":"1-10 TB","count":"~2,000-10,000 orders","volume":"~4 TB; ~2,000-10,000 orders; thousands of analytical runs","years":"2006–2026","yearsOfHistory":"20","growing":"Yes","systems":["Spreadsheets","Inventory system","Email","File storage","Instruments","Accounting"],"contents":["Orders with linked POs, quotes and invoices","Packing lists, freight and customs paperwork","HPLC, GC and IR raw instrument files and reports","Outsourced NMR reports","Certificates of analysis","Safety data sheets and specifications","Hazmat and import/export compliance documents","Bilingual English/Chinese sourcing email (~20 years)","Chemical movement records and user audit log","SOPs","Bookkeeping records"],"formats":["Spreadsheets","Inventory system exports","Instrument raw files","Email","PDF documents"],"delivery":"","labelled":"Partly","labelTypes":["Linked QC results and certificates of analysis per lot"],"personalData":"Low","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"Specialty chemicals custom synthesis & toll manufacturing","tools":[["email","Email"],["accounting","Accounting"],["erp","ERP"],["sheet","Spreadsheets"],["files","File storage"]],"span":"20 years","volumeHead":"~2,000-10,000 orders","sizeTB":4.0,"fromYear":2006,"toYear":2026,"yearsNum":20.0,"tags":["chemistry","specialty chemicals","toll manufacturing","hplc","chromatography","certificates of analysis","sds","supply chain","bilingual email"],"shortTitle":"Specialty chemicals R&D record"},{"ref":"MK-45077","title":"Native engineering files from real survey jobs: 30K+ point-cloud/3D files, CAD/BIM, GIS and mapping rasters (~250 GB)","headline":"Native CAD/BIM (378 files), ~30,500 point-cloud and 3D files, 720 GIS layers and 677 high-resolution mapping rasters from real commercial jobs","description":"Native engineering and mapping files from real commercial survey jobs: CAD/BIM (SketchUp, DWG, DXF, Revit), point clouds and 3D (RCS, LAS/LAZ, PTX, OBJ, PLY, FBX, 3D tiles), GIS layers and high-resolution mapping rasters.","category":"Industrial & sensor","sector":"Industrial & engineering","subsector":"Survey / mapping","dataTypes":["CAD / 3D","Geospatial"],"useCases":["Spatial & 3D models","Engineering agents"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"100 GB-1 TB","count":"~31,000 files","volume":"~250 GB: point cloud/3D 84.5 GB; rasters 165 GB; CAD/BIM 0.4 GB; GIS 0.8 GB","years":"2017–2026","yearsOfHistory":"9","growing":"","systems":[],"contents":[],"formats":["SKP, DWG, DXF, RVT, RCS, LAS, LAZ, PTX, OBJ, PLY, FBX, SHP, GeoJSON, TIFF"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"On request","readiness":"On request","stage":"Inventory ready","scale":"","employees":"","industry":"Survey / mapping","tools":[["design","Revit"]],"span":"9 years","volumeHead":"31K files","sizeTB":0.25,"fromYear":2017,"toYear":2026,"yearsNum":9.0,"tags":["point cloud","cad","bim","gis","survey"],"shortTitle":"Survey point clouds and CAD"},{"ref":"MK-45673","title":"Scientific research and grant-writing environment","headline":"Scientific research and grant-writing agent environment: hypotheses, experiment design, NIH-style proposals and peer-review revision","description":"Environment simulating scientific research development: agents formulate and revise hypotheses and objectives, design experiments, construct methodological arguments, draft NIH-style grant sections, weigh evidence and respond to reviewer critiques over iterative review cycles. Observations mix structured JSON task specs and unstructured scientific text (biomedical context, reference materials, funding requirements, prior drafts, reviewer evaluations). A rubric-based reward scores scientific validity, experimental design, logical consistency, clarity and responsiveness to feedback, penalising unsupported claims and fabricated references. Internal suites include a 72-case beta and an expanded framework of 400+ evaluation objects across seven tracks; difficulty varies by domain, evidence, document length, funding mechanism and number of review cycles. Python/REST, CPU-only scoring, delivered as hosted API access or dedicated deployment.","category":"RL environment / benchmark","sector":"AI & data","subsector":"RL environment: scientific research and grant writing","dataTypes":["RL environment / simulator","Text"],"useCases":["RL environments","Evaluation / benchmarks","Domain fine-tuning"],"origin":"Expert-created","regions":["Not stated"],"languages":["English"],"sizeBand":"","count":"400+ evaluation objects","volume":"","years":"","yearsOfHistory":"","growing":"","systems":["REST API"],"contents":["Scientific research task specifications","Grant-section drafting tasks","Reviewer critiques and revision cycles","Scoring rubrics","400+ evaluation objects across 7 tracks","Baseline comparisons","Integration documentation"],"formats":["JSON","Text"],"delivery":"API","labelled":"Partly","labelTypes":["Scoring rubrics"],"personalData":"None","deidentification":"Not needed","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"RL environment: scientific research and grant writing","tools":[],"span":"","volumeHead":"400 evaluation objects","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["rl environment","scientific reasoning","grant writing","peer review","biomedical","rubric reward","agents"],"shortTitle":"Grant-writing environment"},{"ref":"MK-46289","title":"Asian packaging manufacturer and exporter: email, ERP and quality files","headline":"English-speaking Asian packaging manufacturer: ~12 years of email, ~5.5 years of SAP ERP and ~46K Drive files","description":"Internal data of an English-speaking packaging manufacturer and exporter in Asia, in business for 25+ years. Includes about 12 years of business email, about 5.5 years of SAP ERP records (orders, production, BOMs, quotations) and ~46K Google Drive files (specifications, ISO and quality documents). Around 0.5 TB in total. Email contains staff, customer and supplier contacts. Offered through an intermediary; rights are to be confirmed.","category":"Company operations archive","sector":"Manufacturing","subsector":"Packaging manufacturing and export","dataTypes":["Email","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Computer-use & tool-use agents","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["Asia"],"languages":["English"],"sizeBand":"100 GB-1 TB","count":"~46K Drive files","volume":"~0.5 TB; ~46K Drive files; ~12 years of email; ~5.5 years of ERP records","years":"","yearsOfHistory":"12","growing":"","systems":["Email","ERP (SAP)","File storage (Google Drive)"],"contents":["Business email (~12 years)","ERP orders, production, BOMs and quotations","Product specifications","ISO and quality documents"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Mid-size company","employees":"51–200","industry":"Packaging manufacturing and export","tools":[["googledrive","Google Drive"],["sap","SAP"],["email","Email"]],"span":"12 years","volumeHead":"46K Drive files","sizeTB":0.5,"fromYear":null,"toYear":null,"yearsNum":12.0,"tags":["packaging","manufacturing","export","sap","erp","bom","quality documents","iso"],"shortTitle":"Asian packaging maker ERP archive"},{"ref":"MK-47447","title":"European womenswear brand: full operating archive","headline":"European womenswear brand: ~10 years of email, docs, e-commerce/POS orders, marketing events, support tickets and media","description":"Operating data of a Southern European womenswear brand (e-commerce, owned retail and wholesale), in Spanish and English. Includes ~1.5M stored emails, ~430K Drive documents, ~11K wiki pages, ~200K team-chat events, a ~90K-file corporate knowledge repository, ~200K online orders and ~110K POS orders with ~1.2M stock movements, 3PL orders, returns and delivery proofs, ~170M email-marketing events with campaigns and flows, ~67K support tickets with attachments, ad creatives, invoices, purchases and expenses, and ~290K images and ~30K videos. Also corporate AI work instructions, decision records and verification artefacts that can be assessed for agent training and evaluation. Raw AI chat logs, HR/payroll, credentials and payment identifiers are out of scope.","category":"Company operations archive","sector":"Retail & e-commerce","subsector":"Fashion brand (womenswear, DTC + retail + wholesale)","dataTypes":["Email","Documents","Chat & messaging","Tabular / structured","Image","Video","Text"],"useCases":["Enterprise workflow agents","Computer-use & tool-use agents","Evaluation / benchmarks","Forecasting & time series","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["Europe"],"languages":["Spanish","English"],"sizeBand":"1-10 TB","count":"~1.5M emails","volume":"~1.5M emails; ~430K documents; ~11K wiki pages; ~200K chat events; ~300K online and POS orders; ~1.2M stock movements; ~67K support tickets; ~290K images; ~30K videos","years":"2016–2026","yearsOfHistory":"10","growing":"Yes","systems":["Email","File storage","Wiki","Team chat","E-commerce","POS","Inventory","3PL","Returns","Email marketing","Help desk","Ads","Analytics","Accounting","Expense management","Meeting notes","Knowledge repository"],"contents":["Business email (~1.5M)","Drive documents (~430K)","Wiki pages (~11K)","Team chat events (~200K)","Online and POS orders (~300K)","Stock movements (~1.2M)","3PL orders, returns and delivery proofs","Email-marketing events, campaigns and flows","Support tickets with attachments (~67K)","Ad creatives","Invoices, purchases and expenses","Product images (~290K) and videos (~30K)","Corporate AI instructions, decision records and verification artefacts"],"formats":["Mixed structured and unstructured exports"],"delivery":"","labelled":"Partly","labelTypes":["Decision records","Verification artefacts"],"personalData":"High","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Fashion brand","tools":[["chat","Team chat"],["tickets","Tickets"],["git","Git"],["email","Email"],["wiki","Docs & wiki"],["accounting","Accounting"],["erp","ERP"],["shop","Commerce"],["analytics","Analytics & marketing"],["files","File storage"]],"span":"10 years","volumeHead":"1.5M emails","sizeTB":5.6,"fromYear":2016,"toYear":2026,"yearsNum":10.0,"tags":["fashion","e-commerce","retail","pos","email marketing","customer support","product images","agent evaluation"],"shortTitle":"Womenswear brand archive"},{"ref":"MK-49420","title":"Nonprofit fundraising software startup: strategy, design, docs and all communications","headline":"Nonprofit fundraising-software startup: 4 years of design files, strategy, finance plans, all email and team chat","description":"Four years of a small fundraising-software company serving nonprofits: design files, strategy documents, financial plans, meeting notes and technical documentation in Drive (all company context), every internal and external email conversation, and every internal team-chat conversation. Owner states ~10 TB total.","category":"Code & software","sector":"Software & internet","subsector":"Nonprofit fundraising software (SaaS)","dataTypes":["Documents","Email","Chat & messaging","Image"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"10-100 TB","count":"","volume":"~10 TB","years":"2022–2026","yearsOfHistory":"4","growing":"Yes","systems":["File storage","Email","Team chat","Design files"],"contents":["Design files","Strategy documents","Financial plans","Meeting notes","Technical documentation","Internal and external email","Internal team chat"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"Nonprofit fundraising software","tools":[["chat","Team chat"],["email","Email"],["design","Design files"],["files","File storage"]],"span":"4 years","volumeHead":"10 TB","sizeTB":10.0,"fromYear":2022,"toYear":2026,"yearsNum":4.0,"tags":["saas","nonprofit","fundraising","startup","email","team chat","strategy docs","design files"],"shortTitle":"Fundraising software startup"},{"ref":"MK-50008","title":"UK consumer wellness brand: 10-year operating archive with linked support, orders and policies","headline":"UK wellness brand: 10-year archive linking ~300K support conversations to orders, CRM, accounting and policies","description":"Operating archive (~355 GB gross) of a UK wellness company selling B2C and wholesale across the UK and Europe since 2016. It holds ~298K customer-support conversations (~332K customer/agent messages with internal notes and attachments), ~87K CRM records, ~389K emails from five archived staff inboxes, ~17K orders across consumer and B2B stores with refunds and order notes, ~54K accounting records (invoices, payments, bank transactions) plus ~890K historical ledger lines, ~217K document-store entries and a second file archive, email-marketing contacts and ~520 campaigns, ~6.8K calendar events and team-chat history. Documents include sales playbooks, CRM guidelines, onboarding, SOPs, supplier and customer correspondence, product catalogues, fulfilment trackers, finance workbooks and historical refund/returns/shipping policies, so support cases can be linked to the order and the policy in force at the time.","category":"Company operations archive","sector":"Retail & e-commerce","subsector":"Wellness consumer brand (B2C and wholesale)","dataTypes":["Chat & messaging","Email","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","SFT / instruction tuning","Evaluation / benchmarks","Retrieval / RAG","Domain fine-tuning"],"origin":"First-party company records","regions":["Europe"],"languages":[],"sizeBand":"100 GB-1 TB","count":"~298K support conversations","volume":"~355 GB gross; 297,563 support conversations (332,182 messages); 388,573 emails; 17,386 orders; 87,302 CRM records; 54,439 accounting records (+890K ledger lines); 216,923 document-store files; 523 campaigns","years":"2016–2026","yearsOfHistory":"10","growing":"Yes","systems":["Help desk","CRM","Email","E-commerce","Accounting","File storage","Team chat","Email marketing","Calendar"],"contents":["Customer-support conversations with agent notes (~298K)","CRM people, organisations, deals and notes (~87K)","Staff email (~389K messages, 5 inboxes)","Consumer and B2B orders with refunds (~17K)","Accounting invoices, payments and bank transactions (~54K) + ~890K ledger lines","Document store (~217K entries): playbooks, SOPs, onboarding","Historical refund, returns and shipping policies","Order and fulfilment trackers","Email-marketing contacts and campaigns","Calendar events (~6.8K)","Team chat messages"],"formats":["JSON","CSV","XLSX","MBOX","XML","SQL","Native files"],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"Owner-held","sample":"","readiness":"2-6 weeks","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"Wellness consumer brand","tools":[["chat","Team chat"],["tickets","Tickets"],["email","Email"],["crm","CRM"],["accounting","Accounting"],["shop","Commerce"],["analytics","Analytics & marketing"],["calendar","Calendar"],["files","File storage"]],"span":"10 years","volumeHead":"298K support conversations","sizeTB":0.355,"fromYear":2016,"toYear":2026,"yearsNum":10.0,"tags":["customer support","wellness","ecommerce","crm","email archive","accounting","policies","sops","uk","wholesale"],"shortTitle":"Wellness brand archive"},{"ref":"MK-52402","title":"Replacement-parts manufacturer: commerce, fitment and catalogue data","headline":"Replacement-parts manufacturer: ERP transactions, OEM fitment cross-reference and before/after catalogue copy with sales outcomes","description":"US aftermarket manufacturer of compatible replacement parts for airless paint-spray equipment (spray tips, pump repair kits, guns, accessories), selling through a dealer network, its own direct-to-consumer site and a marketplace. ERP data (2020-2026) covers orders, invoices, POs, inventory, pricing, ~20-25K SKUs and dealer/customer accounts forming a B2B dealer transaction graph; e-commerce (~19K product records, orders with purchasing-entity and ad-attribution attributes) and a legacy store archive; marketplace orders, inventory and ad campaigns; email marketing; web analytics and search queries; reviews, Q&A, loyalty and survey feedback; and email with dealer quotes and fitment questions. Two custom assets stand out: an OEM fitment cross-reference (~43K fitment rows, ~5.5K OEM part mappings covering 15 equipment brands and 1,200+ models) and ~25K SKU descriptions before and after a copy rewrite, paired with measured sales outcomes.","category":"Company operations archive","sector":"Manufacturing","subsector":"Aftermarket spray-equipment parts manufacturer","dataTypes":["Tabular / structured","Text","Email","Documents","Image"],"useCases":["Enterprise workflow agents","Preference / RLHF","Domain fine-tuning","Forecasting & time series","Retrieval / RAG"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"1-10 TB","count":"~20-25K SKUs","volume":"~1-4 TB (estimate); ~20-25K SKUs; ~43K fitment rows; ~5.5K OEM part mappings; ~25K before/after SKU descriptions","years":"2020–2026","yearsOfHistory":"6","growing":"Yes","systems":["ERP","E-commerce","Marketplace seller account","Email marketing","Analytics","Reviews & loyalty","Email","Custom fitment app"],"contents":["ERP orders, invoices and POs","Inventory and pricing (~20-25K SKUs)","Dealer/customer accounts and B2B transaction graph","E-commerce product records (~19K) and orders","Marketplace orders, inventory and ad campaigns","Email marketing campaigns, flows and engagement","Site analytics and search queries","Product reviews, Q&A, loyalty and surveys","OEM fitment cross-reference (~43K rows, ~5.5K OEM mappings)","Before/after SKU descriptions (~25K) with sales outcomes","Dealer quotes, price lists and MAP policy history","Warranty/returns records and SOPs"],"formats":["CSV"],"delivery":"","labelled":"Partly","labelTypes":["Fitment compatibility mappings","Copy rewrites paired with sales outcomes"],"personalData":"Moderate","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"On request / under NDA","readiness":"2-6 weeks","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Aftermarket spray-equipment parts manufacturer","tools":[["email","Email"],["erp","ERP"],["shop","Commerce"],["analytics","Analytics & marketing"]],"span":"6 years","volumeHead":"~20-25K SKUs","sizeTB":null,"fromYear":2020,"toYear":2026,"yearsNum":6.0,"tags":["manufacturing","aftermarket parts","fitment","erp","ecommerce","dealer network","catalog copy","sales outcomes","b2b"],"shortTitle":"Replacement-parts catalogue data"},{"ref":"MK-55639","title":"Computer-use agent lab: trajectories, sessions and research code","headline":"Computer-use agent lab: 1,375 Android trajectory videos with step-level JSON actions, desktop sessions, checkpoints and agent code","description":"Private working record of a small AI lab building computer-use agents for Android, Windows, macOS and Ubuntu. Includes 97 Android agent runs (1,375 screen videos, 1,224 step-by-step JSON action files, 1,256 logs and 45 success recordings, mostly on a public Android benchmark task suite), 98 desktop-agent observer sessions (screens plus actions), and ~422 GB of model checkpoints with CLIP features of real vs AI-generated images from earlier content-provenance research. Also private agent-harness code (planner/reflector prompts), a fast GUI-parsing server and GUI-grounding fine-tuning code, perceptual-hashing and deepfake-detection models, weekly check-in notes, talk decks, grant applications and 7 recorded meeting transcripts. Public models and public repos are excluded. Offered by an authorised reseller.","category":"Agent traces","sector":"AI & data","subsector":"Computer-use AI agent lab","dataTypes":["Agent trajectories","Screen recordings / UI traces","Video","Code","Documents","Text"],"useCases":["Computer-use & tool-use agents","Evaluation / benchmarks","Vision & perception","Coding models","SFT / instruction tuning"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"100 GB-1 TB","count":"1,375 trajectory videos","volume":"~453 GB (422 GB checkpoints); 1,375 trajectory videos + 1,224 action files; 98 desktop sessions; 7 repos","years":"2023–2026","yearsOfHistory":"3","growing":"","systems":["Cloud storage","Git","File storage","Team chat","Meeting recorder","Email"],"contents":["Android agent run videos (1,375)","Step-by-step JSON action files (1,224) and run logs","Task success recordings","Desktop-agent observer sessions (98)","Model checkpoints (~422 GB)","CLIP image features (real vs AI-generated)","Agent harness code with planner/reflector prompts","GUI-grounding model fine-tuning code","Weekly research check-in notes","Talk decks and grant applications","Meeting transcripts (7 meetings, ~6.7 h)"],"formats":["MP4 / video","JSON","Logs","Model checkpoints","CSV","Git","DOCX","PPTX","PDF"],"delivery":"Download","labelled":"Partly","labelTypes":["Step-level agent actions","Task success recordings"],"personalData":"Moderate","deidentification":"","rights":"Via authorised intermediary","sample":"","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"Computer-use AI agent lab","tools":[["chat","Team chat"],["git","Git"],["email","Email"],["calls","Calls"],["files","File storage"]],"span":"3 years","volumeHead":"1.4K trajectory videos","sizeTB":0.453,"fromYear":2023,"toYear":2026,"yearsNum":3.0,"tags":["computer-use","agent trajectories","android","gui agents","screen recordings","model checkpoints","ai lab","action logs"],"shortTitle":"Computer-use agent lab"},{"ref":"MK-56196","title":"Event-structure hire company: 15 years of jobs from quote to build, costing and safety records","headline":"UK temporary-structure hire company: ~1,500 jobs from quote to build sign-off, job costing, site plans, H&S and SOPs; CRM, accounting and 15 years of files (~250 GB)","description":"Hire and installation of marquees and temporary structures. Full job lifecycle: quotes and bookings, load lists, build sign-offs, site plans, job back-costing and margin, sales metrics, health-and-safety records and SOPs. ~34K indexed documents (~20 GB) and ~100 GB of media; CRM (6 years), file storage (15 years), accounting (10 years) and 5 mailboxes back to 2013. Supplier documents removed before delivery.","category":"Company operations archive","sector":"Construction & real estate","subsector":"Event structures hire","dataTypes":["Documents","Email","Image","Tabular / structured"],"useCases":["Enterprise workflow agents","Domain fine-tuning"],"origin":"First-party company records","regions":["Europe"],"languages":[],"sizeBand":"100 GB-1 TB","count":"~1,500 jobs","volume":"~250 GB; ~1,500 jobs; ~34K indexed documents; ~100 GB media","years":"2011–2026","yearsOfHistory":"15","growing":"","systems":["CRM","File storage","Accounting","Email"],"contents":["Quotes and bookings","Load lists","Build sign-offs","Site plans","Job costing","H&S records","SOPs","Site media"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"Planned / on request","rights":"Owner-held","sample":"On request / under NDA","readiness":"On request","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Event structures hire","tools":[["email","Email"],["crm","CRM"],["accounting","Accounting"],["files","File storage"]],"span":"15 years","volumeHead":"1.5K jobs","sizeTB":0.25,"fromYear":2011,"toYear":2026,"yearsNum":15.0,"tags":["events","construction","job costing","sops","uk"],"shortTitle":"Event-structure hire records"},{"ref":"MK-56991","title":"DTC creative and performance marketing agency: 4-year archive","headline":"DTC creative and performance marketing agency: 4 years of CRM, briefs, creative, performance tracking and client chat","description":"Operating data of a creative and performance marketing agency serving direct-to-consumer brands, ~2 TB over about four years: CRM leads, quotes, invoices and payments, internal and external email, internal SOPs, briefs, drafts, guidelines and creative outputs, project files, performance-tracking sheets and charts, P&Ls and balance sheets, team and client chat, project management, signed client contracts, design renders and a creator database.","category":"Company operations archive","sector":"Marketing, media & publishing","subsector":"Creative and performance marketing agency (DTC)","dataTypes":["Email","Documents","Tabular / structured","Image","Chat & messaging"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"1-10 TB","count":"","volume":"~2 TB","years":"2022–2026","yearsOfHistory":"4","growing":"Yes","systems":["CRM","Email","File storage","Spreadsheets","Accounting","Team chat","Project management","E-signature","Design files"],"contents":["CRM leads and quotes","Invoices and payments","Creative briefs and drafts","SOPs and brand guidelines","Creative outputs and design renders","Campaign performance tracking","P&Ls and balance sheets","Internal and client chat","Signed client contracts","Creator database"],"formats":[],"delivery":"","labelled":"Partly","labelTypes":["Campaign performance tracking"],"personalData":"Moderate","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Creative and performance marketing agency","tools":[["chat","Team chat"],["tickets","Tickets"],["email","Email"],["crm","CRM"],["accounting","Accounting"],["design","Design files"],["sheet","Spreadsheets"],["sign","E-signature"],["files","File storage"]],"span":"4 years","volumeHead":"2 TB","sizeTB":2.0,"fromYear":2022,"toYear":2026,"yearsNum":4.0,"tags":["marketing agency","dtc","creative","performance marketing","briefs","crm","creator database","client contracts"],"shortTitle":"DTC creative agency archive"},{"ref":"MK-57952","title":"MMO game studio: full development and operations history","headline":"US MMO game studio: ~1.2M lines of code, ~57K issue-tracker tickets, ~35.7K daily staff work reports, art assets and email, 2013-2025","description":"Full development and operations history of a US studio that built a MMO game. Includes ~1.2M lines of first-party code in Git (~34.8K commits by ~120 authors, 2018-2021, referencing ~14.6K issue-tracker tickets), ~57K issue-tracker issues (2014-2021), ~119 GB of art assets, audio, logs, ~85 design documents, a business and financial archive, and ~125K emails including ~74.7K studio emails with ~35.7K daily staff work reports (2015-2022). Around 225-240 GB in total. Training rights are stated as licensed exclusively to the offering party by the owner; a chain-of-title agreement has been requested. HR records, player and backer email, recruiting threads, possibly privileged legal email and community-chat player data are to be scrubbed or excluded.","category":"Company operations archive","sector":"Media & entertainment","subsector":"Online multiplayer game development","dataTypes":["Code","Email","Documents","Image","Audio / speech","Tabular / structured","Text"],"useCases":["Coding models","Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"100 GB-1 TB","count":"~1.2M lines of code","volume":"~225-240 GB; ~1.2M LOC; ~34.8K commits by ~120 authors; ~57K tickets; ~119 GB art assets; ~85 design docs; ~125K emails incl. ~35.7K daily work reports","years":"2013–2025","yearsOfHistory":"12","growing":"","systems":["Git (GitLab)","Issue tracker (Jira)","Email","Team chat (Discord)","Asset storage","Logs"],"contents":["First-party game code (~1.2M LOC)","Commit history (~34.8K commits)","Issue-tracker tickets (~57K)","Art assets (~119 GB)","Audio","Logs","Design documents (~85)","Business and financial archive","Daily staff work reports (~35.7K)","Company and industry email"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"Planned (scrub or exclude sensitive email and player data)","rights":"Third-party licensed","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Mid-size company","employees":"51–200","industry":"Online multiplayer game development","tools":[["discord","Discord"],["gitlab","GitLab"],["jira","Jira"],["email","Email"]],"span":"12 years","volumeHead":"1.2M lines of code","sizeTB":0.23,"fromYear":2013,"toYear":2025,"yearsNum":12.0,"tags":["game development","mmo","code","commits","issue tracker","art assets","work reports","studio archive"],"shortTitle":"MMO game studio dev history"},{"ref":"MK-58267","title":"Tax-preparation case corpus","headline":"De-identified US tax-preparation case corpus: source documents, workpapers, returns, notices and correspondence across client-years","description":"Real-source individual and business tax cases (1040; 1065, 1120S, 1120) assembled from the archives of several mid-market CPA firms. Each case links source documents (W-2, 1099, 1098, 5498, brokerage trades and basis, bank, mortgage, payroll and retirement records), organizers, workpapers, reconciliations, trial balances, book-to-tax and multistate schedules and K-1s to completed federal and state returns, extensions and diagnostics, together with IRS and state notices and payment follow-up, selected client-preparer email, review notes and firm SOPs for intake, processing, e-filing and collections. Client-years stay connected with consistent synthetic identities; provider and personal identities are substituted. About 0.2 TB now, ~11 years of history, with larger batches available and the corpus expanding weekly.","category":"Expert workflows","sector":"Professional services","subsector":"Tax preparation (CPA firms)","dataTypes":["Documents","Email","Tabular / structured","Text"],"useCases":["Enterprise workflow agents","Evaluation / benchmarks","Domain fine-tuning","SFT / instruction tuning","RL environments"],"origin":"Licensed from third parties","regions":["North America"],"languages":[],"sizeBand":"100 GB-1 TB","count":"~0.2 TB of cases","volume":"~0.2 TB current corpus; larger batches available; expanding weekly","years":"~11 years (varies by case)","yearsOfHistory":"11","growing":"Yes","systems":["Tax preparation software","Document archive","Practice management","Client portal","Spreadsheets","Email","Accounting","Payroll"],"contents":["Individual tax cases with source forms","Business returns (1065, 1120S, 1120) with financials","Workpapers and reconciliations","Completed federal and state returns","Tax-agency notices and payment history","Client-preparer correspondence","Review notes","Firm SOPs and processing procedures","Linked prior-year context","Expert-scored workflow tasks (separate offering)"],"formats":["PDF","ZIP","DOCX","XLSX","EML","CSV","JSON"],"delivery":"","labelled":"Partly","labelTypes":["Review notes","Expert-reviewed reference work and scoring (separate offering)"],"personalData":"Moderate","deidentification":"Done by owner","rights":"Via authorised intermediary","sample":"Available","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"Tax preparation","tools":[["email","Email"],["accounting","Accounting"],["sheet","Spreadsheets"],["hr","HR & payroll"]],"span":"11 years","volumeHead":"0.2 TB of cases","sizeTB":0.2,"fromYear":null,"toYear":null,"yearsNum":11.0,"tags":["tax","accounting","cpa","tax returns","workpapers","de-identified","financial documents","agent evaluation"],"shortTitle":"Tax-preparation cases"},{"ref":"MK-63508","title":"Texas construction cost data: job costs, bids, change orders, quotes and takeoffs","headline":"Texas contractor cost data: job cost reports, estimates, bid sheets, change orders, quotes and takeoffs (2001-2024)","description":"Construction estimating and cost records from Texas contractors spanning 2001-2024: job cost reports, estimates and bid sheets, change orders, subcontractor and supplier quotes, and quantity takeoffs. These are records contractors normally guard closely. For trade-job cost estimation, competitive pricing, bid-docket preparation and change-order processing agents.","category":"Industrial & sensor","sector":"Construction & real estate","subsector":"Construction estimating and job costing","dataTypes":["Documents","Tabular / structured"],"useCases":["Domain fine-tuning","Enterprise workflow agents","Forecasting & time series","Retrieval / RAG"],"origin":"Licensed from third parties","regions":["North America"],"languages":["English"],"sizeBand":"100 GB-1 TB","count":"~150 GB","volume":"~150 GB","years":"2001–2024","yearsOfHistory":"23","growing":"","systems":["Estimating software","File storage"],"contents":["Job cost reports","Estimates and bid sheets","Change orders","Subcontractor quotes","Supplier quotes","Quantity takeoffs"],"formats":["CSV","PDF","JSON","BFX"],"delivery":"Cloud bucket (S3)","labelled":"No","labelTypes":[],"personalData":"None","deidentification":"","rights":"To be confirmed","sample":"On request / under NDA","readiness":"On request","stage":"Inventory ready","scale":"","employees":"","industry":"Construction estimating and job costing","tools":[["design","Design files"],["files","File storage"]],"span":"23 years","volumeHead":"150 GB","sizeTB":0.15,"fromYear":2001,"toYear":2024,"yearsNum":23.0,"tags":["construction","cost estimating","bids","change orders","takeoffs","quotes","texas","job costing"],"shortTitle":"Texas construction cost data"},{"ref":"MK-71325","title":"Industrial seals manufacturer: 23-year records and CAD","headline":"Industrial seals and rubber-components manufacturer: ~20 years of CRM/accounting, client orders, CAD drawings and email","description":"Records of a supplier and manufacturer of seals and rubber components for mining, factory and service companies: CRM/accounting system data (accounting, storage, inventory, client information and orders, in batches 2015-2018 and 2018-2026), company strategy documents and client files on drives, CAD drawings of client orders, and ~11 years of email covering contracts, communications and orders; roughly 1,000+ folders in total. Records from 2003-2015 are mostly on paper, with some digitised.","category":"Company operations archive","sector":"Manufacturing","subsector":"Industrial seals and rubber components","dataTypes":["Tabular / structured","3D / CAD","Documents","Email"],"useCases":["Domain fine-tuning","Enterprise workflow agents","Forecasting & time series"],"origin":"First-party company records","regions":["East Asia"],"languages":[],"sizeBand":"100 GB-1 TB","count":"~1,000+ folders","volume":"~500 GB; ~1,000+ folders","years":"2003–2026","yearsOfHistory":"23","growing":"Yes","systems":["CRM","Accounting","File storage","Email","Paper archive"],"contents":["Accounting records","Inventory and storage records","Client information and orders","CAD drawings of client orders","Company strategy documents","Contracts","Order and client email"],"formats":["CRM export","Files","CAD","Email"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"2-6 weeks","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Industrial seals and rubber components","tools":[["email","Email"],["crm","CRM"],["accounting","Accounting"],["files","File storage"]],"span":"23 years","volumeHead":"1K folders","sizeTB":0.5,"fromYear":2003,"toYear":2026,"yearsNum":23.0,"tags":["manufacturing","seals","rubber components","cad","mining supply","orders","industrial"],"shortTitle":"Industrial seals maker archive"},{"ref":"MK-72951","title":"Coding-education platform: 4 years of repos, graded submissions, community chat and company records (~330 GB)","headline":"Engineering-education platform (100K+ engineer network): 106 repos (~950K lines), 16 databases (~14M rows), ~466K community messages, ~149K emails, 153 GB of submissions and video","description":"Four years of a platform where engineers build projects with AI coaching: 106 repos (~950K source lines), 16 Postgres databases (~295 tables, ~14M rows), ~466K community chat messages, ~12K team chat messages, ~149K emails, 153 GB of participant submissions and video, ~12.5K participant repo links, plus mined research tasks and evaluations. Graded code attempts and AI mock interviews are included. Some sets need clean-up (duplicate repos, dead links).","category":"Code & software","sector":"Education","subsector":"Engineering education","dataTypes":["Code","Chat & messaging","Email","Video","Tabular / structured"],"useCases":["Coding models","Evaluation / benchmarks","SFT / instruction tuning"],"origin":"Platform / user-generated","regions":["North America"],"languages":[],"sizeBand":"100 GB-1 TB","count":"106 repos; ~14M rows","volume":"~330 GB; 106 repos; ~14M DB rows; ~466K chat messages; ~149K emails","years":"2022–2026","yearsOfHistory":"4","growing":"","systems":["Git","Database","Team chat","Email","File storage"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"On request","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Engineering education","tools":[["postgresql","Postgres"],["chat","Team chat"],["git","Git"],["email","Email"],["files","File storage"]],"span":"4 years","volumeHead":"106 repos","sizeTB":0.33,"fromYear":2022,"toYear":2026,"yearsNum":4.0,"tags":["coding","education","graded submissions","interviews","repos"],"shortTitle":"Coding-education platform"},{"ref":"MK-76935","title":"US multifamily property manager: 10 years of AppFolio work orders, bills and invoices plus email and files (~50 GB)","headline":"Multifamily property-management company: 10 years of AppFolio invoices, bills and work orders, SharePoint files and 20 mailboxes","description":"Operating records of a US multifamily property-management company: ten years of property-management system data (invoices, vendor bills, maintenance work orders), shared files and email.","category":"Company operations archive","sector":"Real estate & construction","subsector":"Property management","dataTypes":["Tabular / structured","Email","Documents"],"useCases":["Enterprise workflow agents","Back-office agents"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"10-100 GB","count":"","volume":"~50 GB; 20 mailboxes","years":"2016–2026","yearsOfHistory":"10","growing":"","systems":["AppFolio","SharePoint","Exchange 365"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Medium","deidentification":"","rights":"Owner-held","sample":"","readiness":"Export 1-2 weeks","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Property management","tools":[["files","SharePoint"],["git","Git"]],"span":"10 years","volumeHead":"50 GB","sizeTB":0.05,"fromYear":2016,"toYear":2026,"yearsNum":10.0,"tags":["property management","work orders","appfolio","back office"],"shortTitle":"Multifamily property operations"},{"ref":"MK-77671","title":"Creator and expert marketplace: 5 years of code, CRM, sales calls, email and product docs","headline":"Expert marketplace: 96 repos (~2.6M LOC, 7.6K PRs), 314K CRM contacts with 187K calls and ~83K sales-call recordings, 91 mailboxes, ~200K documents (~2.1 TB)","description":"Marketplace for paid 1:1 expert sessions, webinars and courses (India, North America, UK, Gulf). Code (96 repos, ~2.6M lines, 7,594 PRs), CRM (313,772 contacts, 186,643 calls, 189,457 notes, 97,950 tasks, ~82,789 internal sales-call recordings), 91 mailboxes, ~200K documents including 500-650 PRDs, specs and SOPs, and team chat. Production database and session recordings excluded.","category":"Company operations archive","sector":"","subsector":"","dataTypes":["Code","CRM","Call recordings","Email","Documents","Team chat"],"useCases":["Enterprise / business agents","Sales agents","Coding models"],"origin":"","regions":["Global"],"languages":[],"sizeBand":"","count":"","volume":"~2.1 TB; ~2.8M files; ~1B rows","years":"2021-2026","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held (company-owned data)","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Mid-size company","employees":"51–200","industry":"","tools":[],"span":"","volumeHead":"2.1 TB","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":[],"shortTitle":"Creator marketplace archive"},{"ref":"MK-78223","title":"Commercial and multifamily real estate: operations corpus with scored task outputs (~14 TB)","headline":"Production-derived real estate finance and operations corpus (~14 TB, 3.6M objects): rent rolls, T12s, ledgers, work orders, underwriting models and 30K+ evaluator-scored task outputs","description":"Real estate data from production use: rent rolls, trailing-12 statements, general ledgers and trial balances, work orders, underwriting models, ~527 GB of historical rent and pricing data, and 30K+ task-output pairs with evaluator scores and accepted outputs. Tenant, owner and lender details are removed before delivery.","category":"Financial & transactional","sector":"Construction & real estate","subsector":"Commercial and multifamily real estate","dataTypes":["Tabular / structured","Documents"],"useCases":["Domain fine-tuning","Preference / RLHF","Evaluation / benchmarks","Enterprise workflow agents"],"origin":"Mixed","regions":["North America"],"languages":[],"sizeBand":"10-100 TB","count":"~3.61M objects","volume":"~14.1 TB; ~3.61M objects; 30K+ scored task-output pairs","years":"2021–2026","yearsOfHistory":"5","growing":"","systems":[],"contents":[],"formats":["Parquet","CSV","XLSX","PDF"],"delivery":"","labelled":"Partly","labelTypes":[],"personalData":"Moderate","deidentification":"Done by owner","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"Commercial and multifamily real estate","tools":[],"span":"5 years","volumeHead":"3.6M objects","sizeTB":14.1,"fromYear":2021,"toYear":2026,"yearsNum":5.0,"tags":["real estate","underwriting","rent rolls","scored outputs","rl"],"shortTitle":"Commercial real-estate ops"},{"ref":"MK-78945","title":"Creator storytelling platform: ~1M stories and comics from 300K+ writers, plus messages and 20B+ behaviour events (2021-2026)","headline":"~1M user-written stories and comics from 300K+ writers in 30+ countries (incl. a large Arabic-speaking base), 12M in-app messages and 20B+ behavioural events","description":"Content and usage data from a creator storytelling platform: about a million stories and comics written by 300K+ writers across 30+ countries, 12 million in-app messages and over 20 billion behavioural events. Narrative text is the cleanest first package; messages and events would be a de-identified, scoped subset.","category":"Text & documents","sector":"Media & entertainment","subsector":"Creator platform","dataTypes":["Text","Images","Chat & messaging","Tabular / structured"],"useCases":["Creative writing","Multilingual","Recommendation"],"origin":"Platform / user-generated","regions":["Global"],"languages":[],"sizeBand":"","count":"~1M stories","volume":"~1M stories; 12M messages; 20B+ events (inventory counts)","years":"2021–2026","yearsOfHistory":"5","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Medium","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Timing to confirm","stage":"Inventory ready","scale":"","employees":"","industry":"Creator platform","tools":[],"span":"5 years","volumeHead":"1M stories","sizeTB":null,"fromYear":2021,"toYear":2026,"yearsNum":5.0,"tags":["stories","comics","creative writing","arabic","ugc"],"shortTitle":"Creator storytelling platform"},{"ref":"MK-82188","title":"Copper processing plant engineering: 73,000+ native DWG drawings and project records (~101 GB)","headline":"73,109 files from engineering partners on copper processing plants: native DWG drawings (equipment, pumps, conveyors, electrical, civil) plus PDFs, Word, Excel and MS Project records","description":"Engineering design archives for copper processing plants from a network of Latin American engineering firms: native AutoCAD drawings covering equipment, pumps, conveyors, electrical and civil works, with the PDFs, specifications, spreadsheets and project schedules around them. The owner works with hundreds of engineering firms holding 30K-200K files each, so more volume is available.","category":"Industrial & sensor","sector":"Industrial & engineering","subsector":"Mining and minerals processing","dataTypes":["CAD / 3D","Documents","Tabular / structured"],"useCases":["Engineering agents","Document understanding"],"origin":"First-party company records","regions":["Latin America"],"languages":[],"sizeBand":"100 GB-1 TB","count":"73,109 files","volume":"73,109 files; ~101 GB; sample of 366 files available","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":["DWG, PDF, DOCX, XLSX, MPP"],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"Sample available","readiness":"Ready now (sample); more on request","stage":"Inventory ready","scale":"","employees":"","industry":"Mining and minerals processing","tools":[["sheet","Excel"],["autocad","AutoCAD"]],"span":"","volumeHead":"73K files","sizeTB":0.1,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["cad","dwg","mining","copper","plant engineering"],"shortTitle":"Copper plant engineering drawings"},{"ref":"MK-82489","title":"Browser-based social poker app: codebase and game data","headline":"Real-time browser poker app (play money): ~7-year production codebase, game database, chat and analytics","description":"Free browser-based poker for friends with play-money tables, built-in video chat, table chat, multi-table tournaments and a host-approved buy-in/cash-out ledger. Includes the full private production codebase (~3,050 commits, 324 PRs; React front end, real-time socket game server, tournaments, Kubernetes deployment and tests), a game database with table settings, per-seat stacks and profit/loss, ledgers, hands-won counts, free-text table chat and user feedback (no hand-by-hand histories), compressed database dumps and early CSV/SQLite exports, plus session replays, per-user in-game action events, error events, web analytics, wiki/design docs and a shared team inbox held at vendors. No real-money payments.","category":"Code & software","sector":"Software & internet","subsector":"Online social gaming (play-money poker)","dataTypes":["Code","Tabular / structured","Chat & messaging","Screen recordings / UI traces","Documents","Email"],"useCases":["Coding models","RL environments","World models","Forecasting & time series"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"<10 GB","count":"~3,050 commits","volume":"3.5 GB compressed (est. 5-40 GB); ~3,050 commits / 324 PRs; est. 1M-20M analytics events","years":"2019–2026","yearsOfHistory":"7","growing":"","systems":["Database","Cloud storage","Git","Analytics","Session replay","Error monitoring","Wiki","Email","Design files"],"contents":["Production codebase (~3,050 commits, 324 PRs)","Game records with per-seat stacks and profit/loss","Buy-in/cash-out ledgers","Hands-won counts and tournaments","Free-text table chat","User feedback","Session replays and click events","Per-user in-game action events (est. 1M-20M)","Error events tagged by game","Planning and design docs"],"formats":["Postgres dump","CSV","SQLite","Git","MBOX","HTML / Markdown"],"delivery":"Download","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Via authorised intermediary","sample":"","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Online social gaming","tools":[["git","Git"],["email","Email"],["wiki","Docs & wiki"],["database","Database"],["design","Design files"],["analytics","Analytics & marketing"],["files","File storage"]],"span":"7 years","volumeHead":"3K commits","sizeTB":0.0035,"fromYear":2019,"toYear":2026,"yearsNum":7.0,"tags":["poker","multiplayer game","game logs","codebase","table chat","session replays","analytics events","play money"],"shortTitle":"Social poker app codebase"},{"ref":"MK-83704","title":"Chest and general X-rays with radiologist reports","headline":"~15M chest and general X-rays (DICOM) from Indian diagnostic centres, paired with radiologist reports","description":"About 15 million X-ray images collected across multiple diagnostic centres in India from November 2016 to the present, each paired with a radiologist report. Images are DICOM; there are no image-level annotations. The owner states PII/PHI and on-image identifiers have been removed, consent was obtained and no minors are included. Use is limited to research and AI training (radiology anomaly detection, triage, validation and benchmarking).","category":"Healthcare & life sciences","sector":"Healthcare & life sciences","subsector":"Diagnostic radiology (X-ray)","dataTypes":["Image","Text"],"useCases":["Vision & perception","Domain fine-tuning","Evaluation / benchmarks"],"origin":"First-party company records","regions":["South Asia"],"languages":[],"sizeBand":"","count":"~15M X-ray images","volume":"~15M images","years":"2016–2026","yearsOfHistory":"9","growing":"Yes","systems":["Local disk","Cloud storage"],"contents":["X-ray images (~15M, DICOM)","Paired radiologist reports","Multi-centre coverage","Data from 2016 onward"],"formats":["DICOM"],"delivery":"","labelled":"Partly","labelTypes":["Paired radiologist reports (no image annotations)"],"personalData":"Moderate","deidentification":"Done by owner","rights":"To be confirmed","sample":"Not available","readiness":"2-6 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"Diagnostic radiology","tools":[["files","File storage"]],"span":"9 years","volumeHead":"15M X-ray images","sizeTB":null,"fromYear":2016,"toYear":2026,"yearsNum":9.0,"tags":["radiology","x-ray","medical imaging","dicom","radiologist reports","chest x-ray","india","healthcare"],"shortTitle":"X-rays with radiologist reports"},{"ref":"MK-84127","title":"Wound-down AI workflow startup: engineering and agent archive","headline":"Wound-down AI workflow startup: code, PRs, team chat, issues and ~21K production agent traces with tool and cost metadata","description":"Archive of a US startup (now wound down) that built analytics and payments tools and AI assistants for business workflows and SMS, covering about one year (2025-2026). Includes 24 private repositories with ~1.1K pull requests and commit/review history; team chat (45 channels, ~25K messages, ~11K replies, ~1.2K files); ~760 issues with comments, projects and cycles; ~170 documents, spreadsheets and PDFs plus meeting notes; ~21K production agent traces (~74K events, ~5K with tool calls) carrying prompts, tool calls, memory operations and token/cost/timing metadata; and ~5.1K AI-gateway calls with paired request/response records. Billing and payment records exist but need separate review. Rough total ~20 GB.","category":"Company operations archive","sector":"Software & internet","subsector":"AI workflow / payments startup (wound down)","dataTypes":["Code","Chat & messaging","Agent trajectories","Documents","Email","Text"],"useCases":["Coding models","Computer-use & tool-use agents","Enterprise workflow agents","Evaluation / benchmarks","SFT / instruction tuning"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"10-100 GB","count":"~21K agent traces","volume":"24 repos / 1,124 PRs; ~36K chat messages; 762 issues; 21,059 agent traces / 74,053 events; 5,109 gateway calls","years":"2025–2026","yearsOfHistory":"1","growing":"No","systems":["Git","Team chat","Issue tracker","File storage","Email","Database","AI gateway logs","Payments"],"contents":["Private repositories (24) with ~1.1K PRs and review history","Team chat messages and threads (~36K)","Issues, comments, projects and cycles (~760 issues)","Product specs and technical docs","Production agent traces (~21K traces, ~74K events)","Tool-call, memory, token and cost metadata","AI-gateway request/response pairs (~5K)","Meeting notes","Billing and subscription records (review needed)"],"formats":["Git","JSON","JSONL","CSV","Native document exports"],"delivery":"","labelled":"Partly","labelTypes":["Agent traces with tool, memory, token, cost and timing metadata"],"personalData":"Moderate","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"On request","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"AI workflow / payments startup","tools":[["chat","Team chat"],["tickets","Tickets"],["git","Git"],["email","Email"],["database","Database"],["files","File storage"]],"span":"1 year","volumeHead":"21K agent traces","sizeTB":0.02,"fromYear":2025,"toYear":2026,"yearsNum":1.0,"tags":["agent traces","startup archive","source code","pull requests","team chat","issue tracker","llm gateway logs","tool use","wound down"],"shortTitle":"AI workflow startup archive"},{"ref":"MK-84386","title":"B2B buying-intent dataset: ~21K classified requests with reasoning and human buyer decisions from 120+ founder communities","headline":"~126K messages from private B2B founder communities (Jun-Oct 2026), ~21K requests classified for buying intent with reasoning, ~1.2K human buyer decisions, ~1.2K vendor-match verdicts and an intro-to-qualification outcome chain","description":"Derived, de-identified dataset from a B2B matchmaking service: requests posted in 120+ private founder communities, classified for buying intent with written reasoning; human real-buyer decisions; vendor-match verdicts, some with written reasons; and outcomes from buyer conversations through intros to vendor qualification, with 180+ vendor capability profiles. Delivered as JSON and spreadsheets.","category":"Agent traces","sector":"Software & SaaS","subsector":"B2B marketplace","dataTypes":["Text","Tabular / structured","Labels / judgements"],"useCases":["Sales agents","Reward models","Classification"],"origin":"Platform / user-generated","regions":["Global"],"languages":[],"sizeBand":"","count":"~21K classified requests","volume":"126,394 messages; ~21K classified requests; ~1.2K buyer decisions; ~1.2K match verdicts","years":"Jun-Oct 2026","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"Yes","labelTypes":[],"personalData":"Low","deidentification":"De-identified derived data","rights":"Owner-held, conditions to confirm","sample":"","readiness":"1-2 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"B2B marketplace","tools":[],"span":"","volumeHead":"21K classified requests","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["b2b","buying intent","human judgements","matching","sales"],"shortTitle":"B2B buying-intent requests"},{"ref":"MK-84534","title":"Indian home-construction advisory platform: 6 years of code (~574K LOC, 4.5K reviewed PRs), Slack (~811K msgs), CRM and calls","headline":"Consumer home-construction advisory company: ~4 TB Drive (plans, quote analyses, moodboards, site images), Slack ~811K msgs, CRM ~5.6K advisory deals with outcomes, ~11K calls, code with ~4.5K reviewed PRs","description":"Operating record of a home-construction advisory platform (consumer app plus expert advisors): advisor plans and quote analyses, ~4.8K moodboard PDFs, project images and video, team chat, email, CRM with advisory deals and outcomes, call recordings, and the first-party codebase with commit and code-review history. Non-exclusive licence; owner de-identifies and exports.","category":"Company operations archive","sector":"Real estate & construction","subsector":"Home construction advisory","dataTypes":["Code","Team chat","Email","Documents","Images","Call recordings","Tabular / structured"],"useCases":["Enterprise workflow agents","Coding models","Speech & audio models"],"origin":"First-party company records","regions":["South Asia"],"languages":[],"sizeBand":"1-10 TB","count":"~811K messages; ~4.5K PRs","volume":"Drive ~4.08 TB; Slack ~811K msgs / ~288 channels; email ~82 GB; CRM ~136K contacts, ~5.6K deals; ~688 hrs recordings + ~11.3K calls; ~574K LOC, ~20K commits, ~4.5K PRs","years":"~6 years (CRM 3, calls 1)","yearsOfHistory":"","growing":"","systems":["Google Drive","Slack","Email","CRM","Git"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Medium","deidentification":"Owner-side","rights":"Owner-held","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"Home construction advisory","tools":[["slack","Slack"],["googledrive","Google Drive"],["git","Git"],["email","Email"],["crm","CRM"]],"span":"6 years","volumeHead":"811K messages","sizeTB":4.2,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["construction","consumer","code review","slack","crm","calls"],"shortTitle":"Home-construction advisor codebase"},{"ref":"MK-85436","title":"Classical Arabic books: OCR with every word hand-checked","headline":"Classical Arabic books with every OCR word hand-checked: ~50 books, ~10,000 page images paired with verified text","description":"~50 classical Arabic (Islamic) books, ~10,000 pages, digitised from publicly available scans and OCR'd, then every word checked by hand by paid reviewers for accuracy. Page images are paired with word-level human-verified transcriptions, giving ground truth for Arabic-script OCR and handwriting recognition, which current models handle poorly. Texts span works written from around the 9th century to the present.","category":"Text & documents","sector":"Other","subsector":"Nonprofit Arabic text digitisation / OCR ground truth","dataTypes":["Image","Text","Documents"],"useCases":["Vision & perception","Pre-training","Domain fine-tuning","Evaluation / benchmarks"],"origin":"Mixed","regions":["Not stated"],"languages":["Arabic"],"sizeBand":"","count":"~10,000 pages","volume":"~50 books; ~10,000 pages","years":"800–2026","yearsOfHistory":"","growing":"","systems":["OCR pipeline","Cloud storage"],"contents":["Scanned page images","Word-level hand-verified transcriptions","Classical Arabic religious and scholarly texts","OCR output"],"formats":["PDF","JSON"],"delivery":"Cloud bucket (S3); API","labelled":"Yes","labelTypes":["Word-level human-verified transcription"],"personalData":"None","deidentification":"Not needed","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"Nonprofit Arabic text digitisation / OCR ground truth","tools":[["files","File storage"]],"span":"1226 years","volumeHead":"10K pages","sizeTB":null,"fromYear":800,"toYear":2026,"yearsNum":1226,"tags":["arabic","ocr","handwriting","classical texts","ground truth","transcription","document ai"],"shortTitle":"Hand-checked classical Arabic OCR"},{"ref":"MK-86534","title":"US fashion brand: 5-year design-to-sale archive","headline":"Closed US fashion brand: ~400 styles of patterns, tech packs, supply chain, accounting and sales outcomes linkable by style","description":"Five-year operating archive of a US fashion brand that has stopped trading but remains incorporated. Covers ~400 styles with patterns, tech packs, production and supply-chain records, production and supply-chain email correspondence, accounting and cost records, and sales outcomes, linkable by style from design to sale. The team did not use AI tools, so the material is human-authored. Inventory is preliminary: storage size, exact dates and export methods are still to be confirmed and the exports are not yet joined or standardised.","category":"Company operations archive","sector":"Retail & e-commerce","subsector":"Fashion brand (no longer trading)","dataTypes":["Documents","Email","Tabular / structured","Image","3D / CAD"],"useCases":["Domain fine-tuning","Enterprise workflow agents","Pre-training","Retrieval / RAG"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"","count":"~400 styles","volume":"~400 styles; storage size to be confirmed","years":"~5 years","yearsOfHistory":"5","growing":"No","systems":["Accounting","Email","File storage"],"contents":["Garment patterns","Tech packs","Production and supply-chain records","Production/supply-chain email correspondence","Accounting and cost records","Sales outcomes by style","Style-level linkage (~400 styles)"],"formats":["Existing system exports (not standardised)"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"Fashion brand","tools":[["email","Email"],["accounting","Accounting"],["files","File storage"]],"span":"5 years","volumeHead":"400 styles","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":5.0,"tags":["fashion","apparel","patterns","tech packs","supply chain","design to sale","human-authored","closed company"],"shortTitle":"Fashion brand design-to-sale"},{"ref":"MK-88638","title":"Annotated crop-health field imagery","headline":"Annotated crop-health field imagery: ~75K expert-labelled smartphone photos from ~5K farms, collection ongoing","description":"Field smartphone photos of crops on working farms showing pests, diseases, weeds, nutrient deficiencies and healthy plants, captured to a protocol (close-up, leaf underside, whole leaf, whole plant) by the owner's field teams with farmer permission. About 75,000 images (~150 GB) from ~5,000 farms in several states in north, west and south India, annotated in-house by agriculture graduates with agronomist review. Per-image metadata covers crop, variety, growth stage, district, state, date, season, lighting and view; labels include class with scientific name, boxes and polygons, severity score, confirmation method (visual, lab or soil test) and look-alike flags. Quality steps: near-duplicate removal, rejection of off-protocol photos, double annotation with adjudication and tracked inter-annotator agreement; collection can be directed to specific crops, pests, regions or growth stages.","category":"Geospatial & imagery","sector":"Agriculture","subsector":"Crop health field imagery","dataTypes":["Image","Tabular / structured","Geospatial"],"useCases":["Vision & perception","Domain fine-tuning","Evaluation / benchmarks","Robotics / embodied AI"],"origin":"Purpose-collected (consented capture)","regions":["South Asia"],"languages":["English"],"sizeBand":"100 GB-1 TB","count":"~75,000 annotated images","volume":"~150 GB; ~75,000 annotated images from ~5,000 farms; ongoing","years":"2026","yearsOfHistory":"","growing":"Yes","systems":["Cloud storage"],"contents":["Field crop photos (~75K)","Pest, disease, weed and deficiency class labels with scientific names","Bounding boxes and polygon outlines","Severity scores","Growth stage","Confirmation method (visual, lab, soil test)","Look-alike case flags","District-level location and capture metadata"],"formats":["JPEG","COCO JSON","YOLO","CSV"],"delivery":"Cloud bucket (S3); Hugging Face repository; Secure drive transfer","labelled":"Yes","labelTypes":["Class labels","Bounding boxes","Polygons","Severity scores","Growth stage","Confirmation method","Look-alike flags"],"personalData":"Low","deidentification":"Done by owner","rights":"Owner-held","sample":"Available","readiness":"2-6 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"Crop health field imagery","tools":[["files","File storage"]],"span":"","volumeHead":"75K annotated images","sizeTB":0.15,"fromYear":null,"toYear":2026,"yearsNum":null,"tags":["agriculture","crop disease","pests","plant health","computer vision","annotated images","field imagery","india"],"shortTitle":"Crop-health field imagery"},{"ref":"MK-91052","title":"Wound-down AI software-testing startup: email, team chat, files and recordings (2023-2026)","headline":"AI software-testing startup archive (Europe): ~16K emails, ~38K human team-chat messages, ~3K files incl. financials and ~8.5 GB of recordings","description":"Complete archive of a wound-down startup that built AI software-testing tools: ~16K emails across 4 mailboxes, ~38K human team-chat messages (plus bot messages), ~3K documents, financial records and call recordings. Raw and not yet de-duplicated; NDA-covered material removed.","category":"Company operations archive","sector":"Software & internet","subsector":"AI software testing","dataTypes":["Email","Chat & messaging","Documents","Video"],"useCases":["Enterprise workflow agents","Coding models"],"origin":"First-party company records","regions":["Europe"],"languages":[],"sizeBand":"10-100 GB","count":"~16K emails; ~38K chat messages","volume":"~13.7 GB (~8.5 GB video); ~16K emails; ~38K human chat messages; ~3K files","years":"2023–2026","yearsOfHistory":"3","growing":"","systems":["Email","Team chat","File storage"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"On request / under NDA","readiness":"On request","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"AI software testing","tools":[["chat","Team chat"],["email","Email"],["files","File storage"]],"span":"3 years","volumeHead":"16K emails","sizeTB":0.014,"fromYear":2023,"toYear":2026,"yearsNum":3.0,"tags":["startup","email","slack","ai testing"],"shortTitle":"Software-testing startup archive"},{"ref":"MK-93817","title":"Quality-engineering services firm: ~9 TB of email, files, CRM and accounting since 2012","headline":"IT quality-engineering services firm: ~9 TB of email, file storage, CRM, accounting, Teams and recruiting systems since 2012","description":"Digital quality-engineering (software testing) services company. Email (~3 TB), file storage (~5.5 TB across OneDrive and shared drives), accounting (~0.4 TB), CRM, Teams and recruiting/HR systems, from 2012 to today.","category":"Company operations archive","sector":"Software & internet","subsector":"Software testing services","dataTypes":["Email","Documents","Tabular / structured","Chat & messaging"],"useCases":["Enterprise workflow agents","Coding models","Pre-training"],"origin":"First-party company records","regions":["North America","South Asia"],"languages":[],"sizeBand":"1-10 TB","count":"~9 TB","volume":"~9 TB (email ~3 TB, files ~5.5 TB, accounting ~0.4 TB, CRM ~60 GB, Teams ~56 GB)","years":"2012–2026","yearsOfHistory":"14","growing":"","systems":["Email","File storage","CRM","Accounting","Team chat","Recruiting"],"contents":["Email","Project documents","CRM","Accounting","Teams chat","Recruiting records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"Planned / on request","rights":"Via authorised intermediary","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Large organisation","employees":"200+","industry":"Software testing services","tools":[["files","OneDrive"],["chat","Team chat"],["email","Email"],["crm","CRM"],["accounting","Accounting"],["hr","HR & payroll"]],"span":"14 years","volumeHead":"9 TB","sizeTB":9.0,"fromYear":2012,"toYear":2026,"yearsNum":14.0,"tags":["it services","qa","software testing","email","crm"],"shortTitle":"Quality-engineering firm archive"},{"ref":"MK-97612","title":"US wellness e-commerce brand with training academy: 12-year archive","headline":"US e-commerce and wellness company: 12 years of store orders, finance, contracts, training material, meeting recordings and email","description":"Operating data of a US e-commerce and wellness company that has run its online store for about 12 years and also runs a training academy. Includes financial statements and forecasts, general ledger, invoices and purchase orders, customer orders and subscriptions, contracts, product, supplier, quality and inventory files, marketing assets, training materials, regulatory files, meeting recordings and transcripts, support tickets and email from ~45 inboxes. The inventory sheet gives ~20 TB in total; measured so far: ~60 GB file storage, ~100 GB project management, ~12 GB store data, with meeting recordings and helpdesk still being measured. Contains customer order and payroll personal data.","category":"Company operations archive","sector":"Retail & e-commerce","subsector":"E-commerce wellness brand with training academy","dataTypes":["Tabular / structured","Documents","Email","Chat & messaging","Audio / speech","Video","Image"],"useCases":["Enterprise workflow agents","Computer-use & tool-use agents","Forecasting & time series","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"10-100 TB","count":"~45 email inboxes","volume":"~20 TB total (per inventory sheet); ~60 GB file storage; ~100 GB project management; ~12 GB store data; ~45 inboxes; meeting recordings and helpdesk not yet measured","years":"2014–2026","yearsOfHistory":"12","growing":"","systems":["E-commerce (Shopify)","Accounting (QuickBooks Online)","Email (Gmail)","File storage (Google Drive, Dropbox)","Team chat (Slack)","Meeting recordings (Fathom, Zoom, Google Meet)","Project management (Teamwork)","Email marketing (Klaviyo)","Helpdesk (Gorgias)","Payroll (Gusto, Deel)","Payments (Stripe)","Shipping (ShipStation)","Forms (Google Forms, 123 Formbuilder)","E-signature (DocHub)"],"contents":["Financial statements and forecasts","General ledger","Invoices and POs","Customer orders and subscriptions","Contracts","Product, supplier, quality and inventory files","Marketing assets","Training materials","Regulatory files","Meeting recordings and transcripts","Support tickets","Email"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Export 1-2 weeks","stage":"Inventory ready","scale":"Growing company","employees":"11–50","industry":"E-commerce wellness brand","tools":[["slack","Slack"],["gmail","Gmail"],["googledrive","Google Drive"],["dropbox","Dropbox"],["quickbooks","QuickBooks"],["shopify","Shopify"],["stripe","Stripe"],["zoom","Zoom"],["tickets","Tickets"],["hr","HR & payroll"],["analytics","Analytics & marketing"],["sign","E-signature"]],"span":"12 years","volumeHead":"45 email inboxes","sizeTB":20.0,"fromYear":2014,"toYear":2026,"yearsNum":12.0,"tags":["e-commerce","wellness","subscriptions","training academy","meeting recordings","finance","support tickets","dtc"],"shortTitle":"Wellness e-commerce brand archive"},{"ref":"MK-98580","title":"Startup newsletter (100K+ subscribers): 6-year editorial and partner archive","headline":"Startup newsletter (100K+ subscribers) archive: 6 years of articles, drafts, editorial and partner email, Notion docs","description":"Six-year internal archive of a two-person startup newsletter with 100K+ subscribers covering seed to Series B companies for investors and enthusiasts. Hundreds of published articles with their drafts, thousands of editorial and partner emails, and Notion workspace documents including a three-year partner CRM. The most distinctive material is the drafting and editorial process and partner correspondence; published articles are already public.","category":"Text & documents","sector":"Marketing, media & publishing","subsector":"Startup / venture newsletter","dataTypes":["Text","Email","Documents"],"useCases":["SFT / instruction tuning","Domain fine-tuning","Pre-training","Enterprise workflow agents"],"origin":"First-party company records","regions":["North America"],"languages":["English"],"sizeBand":"<10 GB","count":"Hundreds of articles; thousands of emails","volume":"~5 GB","years":"6 years","yearsOfHistory":"6","growing":"","systems":["Email","Notion (CRM)","Newsletter platform","File storage","Team chat"],"contents":["Published articles","Article drafts","Editorial and partner email threads","Notion documents","Partner CRM records"],"formats":[],"delivery":"","labelled":"No","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"Small team","employees":"1–10","industry":"Startup / venture newsletter","tools":[["notion","Notion"],["chat","Team chat"],["email","Email"],["crm","CRM"],["files","File storage"]],"span":"6 years","volumeHead":"Hundreds of articles","sizeTB":0.005,"fromYear":null,"toYear":null,"yearsNum":6.0,"tags":["newsletter","startups","venture capital","editorial drafts","email","notion","writing"],"shortTitle":"Startup newsletter archive"},{"ref":"MK-98880","title":"Legacy-language toolchain: ~3.9M lines of parser, grammar and rewrite code covering ~90 legacy languages","headline":"Program-transformation company (since 1995): ~3.9M lines across ~7K files of its own toolkit code, grammars and rewrite rules, with parsers for ~90 languages incl. COBOL dialects, PL/I, JCL, RPG, MUMPS","description":"Codebase of a company that has built program analysis and transformation tools since 1995. Core scope: ~7K files and ~3.9M lines: ~2.26M lines in its own parallel programming language, ~700K lines of rewrite specifications, ~660K lines of language grammars (~200 files), ~230K lines of lexer definitions. Parsers and transformation tools cover ~90 languages, including many COBOL dialects, PL/I, JCL, CICS BMS, HLASM, RPG, Natural, MUMPS, JOVIAL, TAL/TACL, VAX Macro, PowerBuilder, Progress 4GL, PLC ladder logic, Verilog and VHDL. A wider tree (~75M lines) exists for reference but includes customer-owned example code. Stored in SVN (no pull requests).","category":"Code & software","sector":"Software & internet","subsector":"Legacy code modernisation","dataTypes":["Code"],"useCases":["Coding models","Pre-training","Domain fine-tuning"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"","count":"~3.9M lines","volume":"Core: ~7K files, ~3.9M lines; wider tree ~75M lines (reference only)","years":"1995–2026","yearsOfHistory":"31","growing":"","systems":[],"contents":[],"formats":["SVN repository"],"delivery":"","labelled":"","labelTypes":[],"personalData":"None","deidentification":"Not needed","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"Legacy code modernisation","tools":[],"span":"31 years","volumeHead":"3.9M lines","sizeTB":null,"fromYear":1995,"toYear":2026,"yearsNum":31.0,"tags":["cobol","legacy code","parsers","grammars","program transformation","pl/i","rpg"],"shortTitle":"Legacy-language toolchain"},{"ref":"MK-99622","title":"Biological reasoning task generators","headline":"Biological reasoning task generators: organ-system and cell-model simulations plus clinical-trial outcome prediction, verifiable rewards","description":"Three task generators for training and evaluating causal biological reasoning. An organ-systems simulator (~10K model coordinates) and a coupled kinetic cell model (~65K states, ~100K fluxes) generate perturbation questions with adjustable difficulty (depth, number of perturbations and targets, opposing interactions), verified by numerical integration against a matched control. A clinical-trial generator asks whether anonymised Phase 2/3 small-molecule trials succeeded, given genetics, target biology, expression, ADMET predictions, binding and a safety panel via an assays tool server, scored against curated outcomes (reward = 1 - Brier) from thousands of labelled trials. Delivered as containerised task packages with graders and agent traces; proposed throughput ~500 simulation tasks and ~200 clinical tasks per week, with a clinician-validated cell-model pilot.","category":"RL environment / benchmark","sector":"Healthcare & life sciences","subsector":"Biology simulation / clinical-trial prediction tasks","dataTypes":["RL environment / simulator","Text","Agent trajectories","Tabular / structured"],"useCases":["RL environments","Evaluation / benchmarks","Domain fine-tuning"],"origin":"Generated / simulated","regions":["Not stated"],"languages":["English"],"sizeBand":"","count":"~500 simulation + ~200 clinical tasks per week","volume":"Organ-system tasks ~500/week; clinical tasks ~200/week; cell-model pilot 5-10","years":"2026","yearsOfHistory":"","growing":"Yes","systems":["Simulation infrastructure","Docker containers"],"contents":["Organ-system perturbation tasks","Cell kinetic-model perturbation tasks","Clinical-trial outcome prediction tasks","Curated trial outcome labels","Private reference answers and graders","Agent reasoning traces","Benchmark results for frontier models","Difficulty controls"],"formats":["Harbor task packages","Graders","Traces"],"delivery":"Buyer-hosted container","labelled":"Yes","labelTypes":["Reference outcomes","Trial success/failure outcomes","Benchmark scores"],"personalData":"None","deidentification":"Done by owner","rights":"To be confirmed","sample":"Available","readiness":"Within 2 weeks","stage":"Inventory ready","scale":"","employees":"","industry":"Biology simulation / clinical-trial prediction tasks","tools":[["docker","Docker"]],"span":"","volumeHead":"500 simulation + ~200 clinical tasks per week","sizeTB":null,"fromYear":2026,"toYear":2026,"yearsNum":null,"tags":["biology","physiology simulation","cell model","clinical trials","drug discovery","task generator","verifiable rewards","reasoning"],"shortTitle":"Biological reasoning tasks"},{"ref":"MK-99855","title":"B2B services sales funnel: attributed leads, conversations and won deals","headline":"B2B services sales funnel: ~11.5K attributed leads joined from ad spend to WhatsApp/email conversations, won deals and payments","description":"Sales and marketing data of a software development services firm (19 years old) that sells to US and international buyers. Over 11,500 leads carry source attribution, industry, product interest, geography and quality rating, with ~23K timestamped stage transitions, joined end to end from ad spend and campaigns through clicks, conversations and meetings to won deals and cash collected: ~100K CRM communications, ~11.5K WhatsApp sales messages, ~27.6K rep activities, ~20.9K notes, 140 campaigns with paid-media performance, 400 won deals and 191 payment milestones. The firm also lists meeting recordings and transcripts, proctoring and assessment recordings, lecture video, codebases and internal business data across ~8 TB of stores.","category":"Financial & transactional","sector":"Software & internet","subsector":"IT development services firm","dataTypes":["Chat & messaging","Email","Tabular / structured","Video","Audio / speech","Code"],"useCases":["Enterprise workflow agents","Preference / RLHF","Forecasting & time series","Domain fine-tuning","SFT / instruction tuning"],"origin":"First-party company records","regions":["South Asia"],"languages":["English"],"sizeBand":"1-10 TB","count":">11,500 leads","volume":"11,500+ leads; 22,957 stage transitions; 100,000 CRM communications; 11,453 WhatsApp messages; 400 won deals; 191 payment milestones; 27,597 activities; 140 campaigns; ~8 TB total stores","years":"","yearsOfHistory":"","growing":"Yes","systems":["CRM","Messaging (WhatsApp)","Email","Ad platforms","Database","Cloud storage","Git","Telephony","Meeting intelligence"],"contents":["Attributed leads (>11.5K)","Stage transitions (~23K)","CRM communications (~100K)","WhatsApp sales messages (~11.5K)","Rep activities (~27.6K) and notes (~20.9K)","Campaigns (140) with paid-media spend and conversions","Won deals (400) and payment milestones (191)","Meeting recordings, transcripts and call scoring","Proctoring and assessment recordings","Lecture video and learning content","Codebases"],"formats":["CRM exports","Database exports"],"delivery":"","labelled":"Partly","labelTypes":["Deal outcomes","Stage transitions","Campaign attribution","Lead quality ratings"],"personalData":"High","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Inventory ready","scale":"Large organisation","employees":"200+","industry":"IT development services firm","tools":[["whatsapp","WhatsApp"],["git","Git"],["email","Email"],["crm","CRM"],["calls","Calls"],["database","Database"],["analytics","Analytics & marketing"],["files","File storage"]],"span":"","volumeHead":">11,500 leads","sizeTB":8.0,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["sales","crm","whatsapp","lead attribution","b2b","conversations","won deals","marketing","india"],"shortTitle":"B2B sales funnel data"},{"ref":"MK-11541","title":"Off-the-shelf ML engineering and scientific task suites","headline":"Off-the-shelf task suites: ML engineering, scientific ML, post-training, autoresearch and terminal science","description":"Ready-made task datasets and environments in five coverage areas, each with more than 200 tasks: ML engineering (serving engines, kernel optimisation, mechanistic interpretability, core ML libraries), scientific ML, post-training, open-ended research optimisation, and terminal-based science tasks (materials science, neuroscience, genomics, formal mathematics, shock physics). The seller already supplies large labs and is extending the off-the-shelf sets to other labs. Licence terms and delivery format not yet stated.","category":"RL environment / benchmark","sector":"AI & data","subsector":"Agent task environments for ML and science","dataTypes":["Code","RL environment / simulator","Text"],"useCases":["RL environments","Evaluation / benchmarks","Coding models","Computer-use & tool-use agents"],"origin":"Expert-created","regions":["Not stated"],"languages":[],"sizeBand":"","count":"200+ tasks per area (5 areas)","volume":"200+ tasks per area","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":["ML serving-engine tasks","Kernel optimisation tasks","Mechanistic interpretability tasks","Core ML library tasks","Scientific ML tasks","Post-training tasks","Open-ended research optimisation tasks","Terminal science tasks (materials, neuroscience, genomics, formal maths, shock physics)"],"formats":["Tasks / environments"],"delivery":"","labelled":"","labelTypes":[],"personalData":"None","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Ready now","stage":"Summary available","scale":"","employees":"","industry":"Agent task environments","tools":[],"span":"","volumeHead":"200 tasks per area","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["ml engineering","scientific ml","agent tasks","benchmarks","kernels","interpretability","terminal","reinforcement learning"],"shortTitle":"ML and science task suites"},{"ref":"MK-14436","title":"Egocentric worker video with synced IMU: construction trades, pharma packaging, cleaning, assembly (500-800 hrs/month)","headline":"Head-mounted 1080p60 video with synced IMU, exposure and lens calibration of workers doing rebar, formwork, drywall, pharma sorting and packing, commercial cleaning and furniture assembly; task-labelled","description":"Egocentric RGB video (1080p60, HEVC, no audio) with synchronised IMU, exposure metadata and lens calibration, captured from real workers on construction sites, pharmaceutical sorting and packaging lines, commercial cleaning and furniture assembly. Delivered in task-labelled batches; can be paired with 3D site scans.","category":"Physical-world video","sector":"Industrial & engineering","subsector":"Egocentric capture","dataTypes":["Video","Sensor / IMU"],"useCases":["Robotics & world models"],"origin":"Collected to order","regions":["Europe"],"languages":[],"sizeBand":"","count":"","volume":"500-800 hours per month","years":"2026 onward","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":["MP4, Parquet, JSON via cloud storage"],"delivery":"","labelled":"Yes","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held","sample":"Sample available","readiness":"Ready now","stage":"Collected to order","scale":"","employees":"","industry":"Egocentric capture","tools":[],"span":"","volumeHead":"500-800 hours per month","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["egocentric","imu","construction","robotics","manipulation"],"shortTitle":"Worker video with synced IMU"},{"ref":"MK-25244","title":"Codebase-search startup archive: 46 repos, 1.5K issues, product docs, mail and real user queries","headline":"Closed enterprise codebase-search startup: 46 repos (1,032 PRs), 1,520 tracked issues, ~250 product and founder docs, ~12K emails, 4K prompt/answer records and ~9K real user queries (~25 GB)","description":"Archive of a venture-backed startup that built an enterprise codebase search and understanding engine (2023-2024). Code: 46 repos (44 private) with 1,032 pull requests; issue tracker: 1,520 issues across 27 projects; wiki: ~105 pages, 70 product docs and 77 founder docs; ~12.3K emails (founder mailboxes); 4,000 prompt/answer records across 1,222 conversations and a 9,199-row query log from real use.","category":"Code & software","sector":"Software & internet","subsector":"Code search / developer tools","dataTypes":["Code","Text","Email","Documents"],"useCases":["Coding models","Retrieval / RAG","Computer-use & tool-use agents"],"origin":"First-party company records","regions":["North America","South Asia"],"languages":[],"sizeBand":"10-100 GB","count":"46 repos","volume":"~25 GB; 46 repos; 1,032 PRs; 1,520 issues; 4K prompt/answer records; ~9.2K query-log rows","years":"2023–2024","yearsOfHistory":"1","growing":"","systems":["Git","Issue tracker","Wiki","Email","File storage"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Ready now","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Code search / developer tools","tools":[["tickets","Tickets"],["git","Git"],["email","Email"],["wiki","Docs & wiki"],["files","File storage"]],"span":"1 year","volumeHead":"46 repos","sizeTB":0.025,"fromYear":2023,"toYear":2024,"yearsNum":1.0,"tags":["code search","developer tools","repos","user queries","prompts"],"shortTitle":"Codebase-search startup archive"},{"ref":"MK-30908","title":"B2B gold and diamond jewellery marketplace: full ERP database","headline":"B2B gold and diamond jewellery marketplace: full ERP database and operations files","description":"Two years of operating records from a B2B jewellery marketplace: the full ERP database (purchase orders and receipts, stock entries, ledgers, quotations, sales orders, delivery notes, invoices) plus company email and shared-drive operations files.","category":"Company operations archive","sector":"Retail & e-commerce","subsector":"B2B jewellery marketplace","dataTypes":["Tabular / structured","Email","Documents"],"useCases":["Enterprise workflow agents","Domain fine-tuning"],"origin":"First-party company records","regions":["South Asia"],"languages":["English"],"sizeBand":"100 GB-1 TB","count":"~300 GB","volume":"~300 GB","years":"2024–2026","yearsOfHistory":"2","growing":"Yes","systems":["ERP","Email","File storage"],"contents":["Purchase orders and receipts","Stock entries","Ledgers","Sales orders and invoices","Delivery notes"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"Ready now","stage":"Summary available","scale":"Mid-size company","employees":"51–200","industry":"B2B jewellery marketplace","tools":[["email","Email"],["erp","ERP"],["files","File storage"]],"span":"2 years","volumeHead":"300 GB","sizeTB":0.3,"fromYear":2024,"toYear":2026,"yearsNum":2.0,"tags":["erp","jewellery","b2b marketplace","gold","inventory","india"],"shortTitle":"Jewellery marketplace ERP"},{"ref":"MK-34365","title":"Engineering RL environments: CAD, PCB/chip design and formal proofs (12K+ tasks)","headline":"12K+ verifiable engineering RL tasks: CAD, PCB and chip design in real tools, plus formal proofs","description":"RL environments built from real, verified electronics and mechanical designs: ~3,400 computer-use engineering tasks in open-source CAD and PCB tools, ~9,400 computer-use chip-design tasks and ~1,500 formal proof tasks. Agents act through shell commands, code edits, program execution and GUI actions with screenshots; final artifacts are graded by deterministic software oracles with binary or partial-credit rewards. Tasks run in self-contained containers and can run in parallel; baselines are available.","category":"RL environment / benchmark","sector":"AI & data","subsector":"Engineering RL environments","dataTypes":["RL environment / simulator","Screen recordings / UI traces","3D / CAD","Code"],"useCases":["RL environments","Computer-use & tool-use agents","Evaluation / benchmarks"],"origin":"Expert-created","regions":["Not stated"],"languages":["English"],"sizeBand":"","count":"12K+ tasks","volume":"12K+ tasks","years":"Not applicable","yearsOfHistory":"","growing":"","systems":[],"contents":["CAD modelling tasks","Schematic and PCB layout tasks","Chip-design tasks","Formal proof tasks","Deterministic verification oracles"],"formats":["Containers"],"delivery":"Buyer-hosted container","labelled":"Yes","labelTypes":["Deterministic verifiers"],"personalData":"None","deidentification":"Not needed","rights":"Owner-held","sample":"On request / under NDA","readiness":"Ready now","stage":"Summary available","scale":"","employees":"","industry":"Engineering RL environments","tools":[],"span":"","volumeHead":"12K tasks","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["rl environment","computer use","cad","pcb","chip design","formal proofs","lean","engineering agents"],"shortTitle":"Engineering RL environments"},{"ref":"MK-36756","title":"American football play-calling RL environment","headline":"American football play-calling RL environment with a LLM coach and calibrated game outcomes","description":"A simulated professional American football game in which a LLM coach chooses plays (run, pass, kick, punt, kneel, spike, timeout) from game-state variables, with outcomes that roughly match real league play. Includes a counterfactual league-average coach model and the seller's own play-by-play pipeline plus expected-points and win-probability models. The environment is built on historical play-by-play data; the seller states the whole stack has been rewritten in-house. Version 2 (defensive formations, coverages, pass types, coach tendencies) is in development, and an interactive demo exists.","category":"RL environment / benchmark","sector":"Other","subsector":"Sports analytics (American football)","dataTypes":["RL environment / simulator","Tabular / structured"],"useCases":["RL environments","Evaluation / benchmarks"],"origin":"Generated / simulated","regions":["North America"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"Yes","systems":[],"contents":["Play-calling action space (run, pass, kick, punt, kneel, spike, timeout)","Game-state variables","Counterfactual league-average coach model","Play-by-play data pipeline","Expected points (EPA) model","Win probability (WPA) model","v2: defensive formations, coverages, pass types, coach tendencies"],"formats":["RL environment"],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Ready now","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Sports analytics","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["reinforcement learning","sports","american football","simulation","decision making","play calling","agents"],"shortTitle":"Football play-calling environment"},{"ref":"MK-38238","title":"AI-agent platform startup: 2 years of sales, investor and product records","headline":"AI-agent platform startup: 2 years of sales, investor, product and call-transcript records","description":"Records of a startup building a no-code business AI-agent platform: customer and investor correspondence, sales, marketing, product and investor decks (many versions), P&L and business-model sheets, call transcripts, product roadmap and planning.","category":"Company operations archive","sector":"Software & internet","subsector":"No-code AI agent platform","dataTypes":["Email","Documents","Text"],"useCases":["Enterprise workflow agents","Domain fine-tuning"],"origin":"First-party company records","regions":["North America"],"languages":["English"],"sizeBand":"10-100 GB","count":"~16 GB","volume":"~16 GB","years":"2024–2026","yearsOfHistory":"2","growing":"Yes","systems":["Email","File storage","Team chat"],"contents":["Customer correspondence","Investor decks and correspondence","Call transcripts","P&L sheets","Product roadmap"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"Ready now","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"No-code AI agent platform","tools":[["chat","Team chat"],["email","Email"],["files","File storage"]],"span":"2 years","volumeHead":"16 GB","sizeTB":0.016,"fromYear":2024,"toYear":2026,"yearsNum":2.0,"tags":["startup","ai agents","sales","investor","saas"],"shortTitle":"AI-agent platform records"},{"ref":"MK-40166","title":"Indian state real-estate regulatory register: ~10K projects with filings, cleaned and linked","headline":"~10K Indian real-estate regulatory filings, cleaned, entity-resolved and tracked daily","description":"Every project on an Indian state's real-estate regulator register (~10,000 projects across 31 districts), cleaned and normalised: status, dates and extensions, ~6,400 resolved promoter identities, district and locations from filed coordinates. ~9,000 projects include full filings: promoters and agents, land, unit-level inventory, civil works, approvals, uploaded documents, complaints, litigation and quarterly progress. Refreshed daily with change history. Compiled from a public government register; personal identifiers in filings are removed before delivery.","category":"Geospatial & imagery","sector":"Construction & real estate","subsector":"Real-estate regulatory filings","dataTypes":["Tabular / structured","Documents","Geospatial"],"useCases":["Domain fine-tuning","Retrieval / RAG","Forecasting & time series","Evaluation / benchmarks"],"origin":"Public / scraped sources","regions":["South Asia"],"languages":["English","Kannada"],"sizeBand":"<10 GB","count":"~10K projects","volume":"~10K projects; ~30M table cells; ~4 GB raw HTML","years":"2017–2026","yearsOfHistory":"9","growing":"Yes","systems":[],"contents":["Project register","Promoter entity resolution","Unit-level inventory","Quarterly progress updates","Complaints and litigation","Approvals and documents","Daily change history"],"formats":["JSON","CSV","SQLite","HTML"],"delivery":"Download; API","labelled":"Yes","labelTypes":["Normalised fields"],"personalData":"Moderate","deidentification":"Planned / on request","rights":"To be confirmed","sample":"Available","readiness":"Ready now","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Real-estate regulatory filings","tools":[],"span":"9 years","volumeHead":"10K projects","sizeTB":0.004,"fromYear":2017,"toYear":2026,"yearsNum":9.0,"tags":["real estate","rera","regulatory filings","india","entity resolution","construction","document qa"],"shortTitle":"Indian real-estate register"},{"ref":"MK-42847","title":"US AI agent-building startup: 2-year operations archive","headline":"US platform for building virtual AI teammates: ~2 TB of email, team chat, code, CRM and files over 2 years","description":"Operating data of a small US startup building virtual AI teammates (an agent-building platform). About 2 TB over 2 years across email, team chat, code repositories, CRM, file storage and HR. The submission gives little detail on contents; whether any customer data is included is not yet known.","category":"Company operations archive","sector":"Software & internet","subsector":"AI agent-building platform","dataTypes":["Email","Chat & messaging","Code","Documents"],"useCases":["Enterprise workflow agents","Coding models"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"1-10 TB","count":"~2 TB","volume":"~2 TB over 2 years","years":"","yearsOfHistory":"2","growing":"","systems":["Email (Gmail)","Team chat (Slack)","Git (GitHub)","CRM","File storage (Google Drive)","HR"],"contents":["Email","Team chat","Code repositories","CRM records","Drive files","HR records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Ready now","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"AI agent-building platform","tools":[["slack","Slack"],["github","GitHub"],["gmail","Gmail"],["googledrive","Google Drive"],["crm","CRM"],["hr","HR & payroll"]],"span":"2 years","volumeHead":"2 TB","sizeTB":2.0,"fromYear":null,"toYear":null,"yearsNum":2.0,"tags":["ai agents","startup","slack","github","email","crm"],"shortTitle":"AI-teammate startup ops archive"},{"ref":"MK-46495","title":"Mexican legal and medical case studies answered by three credentialed experts each (Spanish)","headline":"Mexican legal and medical case studies, each answered by three credentialed experts with reasoning and agreement scores (Spanish)","description":"Mexican legal case studies written from scratch by practising, certified Mexican lawyers, based on real cases and anonymised; 1,000+ cases written, 128 answered by three experts so far, with further annotation under way. Intended to improve legal reasoning and answers under Mexican law. The owner can also produce Mexican medicine and accounting cases, custom practice areas and monthly runs. A 10-case sample (9 medicine across epidemiology, family and general medicine; 1 corporate law) shows each case with scenario, Mexican context, question, three independent expert answers with reasoning and limitations citing Mexican clinical guidelines and norms, a difficulty level and an inter-expert agreement score (84-95/100).","category":"Text & documents","sector":"Legal","subsector":"Mexican law case studies","dataTypes":["Documents","Text"],"useCases":["SFT / instruction tuning","Domain fine-tuning","Evaluation / benchmarks"],"origin":"Expert-created","regions":["Latin America"],"languages":["Spanish"],"sizeBand":"","count":"1,000+ cases","volume":"1,000+ cases written; 128 answered by three experts so far","years":"2026","yearsOfHistory":"","growing":"Yes","systems":[],"contents":["Expert-written legal case studies","Expert answers per case","Expert review","Mexican law coverage"],"formats":["PDF"],"delivery":"Cloud bucket; API; Download","labelled":"Yes","labelTypes":["Three expert answers per case","Expert reasoning","Inter-expert agreement score","Difficulty level"],"personalData":"Low","deidentification":"","rights":"Owner-held (contributors assign rights)","sample":"Available","readiness":"Ready now","stage":"Summary available","scale":"","employees":"","industry":"Mexican law case studies","tools":[],"span":"","volumeHead":"1K cases","sizeTB":null,"fromYear":2026,"toYear":2026,"yearsNum":null,"tags":["legal","mexico","spanish","case studies","expert answers","legal reasoning"],"shortTitle":"Spanish legal and medical cases"},{"ref":"MK-48748","title":"Math tutoring: problems, worked solutions, handwriting and whiteboard sessions","headline":"AI whiteboard math tutor data: problems, worked solutions, handwriting and tutoring sessions (~20 GB)","description":"Data from an AI visual whiteboard math tutor: math problems from arithmetic to calculus, worked solutions with step-by-step explanations, handwriting data, whiteboard-based tutoring interactions, and AI-generated hints and problem-solving conversations. About 20 GB covering roughly one year, held in a database and file storage. The owner is still inventorying counts, formats and topic coverage, and says student interactions would be cleared and de-identified before sharing.","category":"Text & documents","sector":"Education","subsector":"AI math tutoring","dataTypes":["Text","Image","Chat & messaging"],"useCases":["SFT / instruction tuning","Domain fine-tuning","Vision & perception","Evaluation / benchmarks"],"origin":"Platform / user-generated","regions":["North America"],"languages":[],"sizeBand":"10-100 GB","count":"~20 GB","volume":"~20 GB","years":"1 year","yearsOfHistory":"1","growing":"Yes","systems":["Database","File storage"],"contents":["Math problems (arithmetic to calculus)","Worked solutions","Step-by-step explanations","Handwriting data","Whiteboard tutoring interactions","AI-generated hints","Problem-solving conversations"],"formats":[],"delivery":"","labelled":"Partly","labelTypes":["Worked solutions","Step-by-step explanations"],"personalData":"Moderate","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Ready now","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"AI math tutoring","tools":[["database","Database"],["files","File storage"]],"span":"1 year","volumeHead":"20 GB","sizeTB":0.02,"fromYear":null,"toYear":null,"yearsNum":1.0,"tags":["education","math","tutoring","handwriting","whiteboard","worked solutions","edtech"],"shortTitle":"Math tutoring sessions"},{"ref":"MK-50928","title":"Financial models and accounting workbooks (Excel)","headline":"~500 MB library of Excel financial models, analyses and accounting workbooks (2005-2026)","description":"A US finance consultant's library of Excel financial models, financial analyses and accounting workbooks spanning 2005-2026 (about 500 MB of XLSX files), suited to training models on financial modelling, spreadsheet formatting and general Excel analysis; the seller highlights the underlying analysis and know-how embedded in the models. Files are a mix of the owner's own work and publicly available or free-to-use material, so rights are to be confirmed per file. Partly labelled; a sample is available and files can be delivered by download or shared drive now.","category":"Expert workflows","sector":"Financial services & insurance","subsector":"Financial modelling / accounting consultancy","dataTypes":["Tabular / structured","Documents"],"useCases":["SFT / instruction tuning","Domain fine-tuning","Computer-use & tool-use agents","Evaluation / benchmarks"],"origin":"Mixed","regions":["North America"],"languages":["English"],"sizeBand":"<10 GB","count":"~500 MB of workbooks","volume":"~500 MB","years":"2005–2026","yearsOfHistory":"21","growing":"","systems":["Spreadsheets","File storage"],"contents":["Excel financial models","Financial analyses","Accounting workbooks","Spreadsheet formatting examples"],"formats":["XLSX"],"delivery":"Download; Shared drive","labelled":"Partly","labelTypes":[],"personalData":"None","deidentification":"","rights":"To be confirmed","sample":"Available","readiness":"Ready now","stage":"Summary available","scale":"","employees":"","industry":"Financial modelling / accounting consultancy","tools":[["sheet","Excel"],["files","File storage"]],"span":"21 years","volumeHead":"500 MB of workbooks","sizeTB":0.0005,"fromYear":2005,"toYear":2026,"yearsNum":21.0,"tags":["excel","financial models","accounting","spreadsheets","finance","xlsx","analysis"],"shortTitle":"Financial model workbooks"},{"ref":"MK-51895","title":"Small marketing agency: email, WhatsApp, Slack and Drive records","headline":"Small marketing agency offering 2-4 years of email, WhatsApp, Slack and Google Drive records","description":"Operating data of a small marketing agency covering 2-4 years: Gmail, WhatsApp, Slack and Google Drive files. Volumes not yet stated.","category":"Company operations archive","sector":"Marketing, media & publishing","subsector":"Marketing agency","dataTypes":["Email","Chat & messaging","Documents"],"useCases":["Enterprise workflow agents","Domain fine-tuning"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"2-4","growing":"","systems":["Email (Gmail)","WhatsApp","Team chat (Slack)","File storage (Google Drive)"],"contents":["Email","WhatsApp chats","Team chat","Drive files"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Ready now","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Marketing agency","tools":[["slack","Slack"],["whatsapp","WhatsApp"],["gmail","Gmail"],["googledrive","Google Drive"],["hr","HR & payroll"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["marketing agency","whatsapp","slack","email","google drive"],"shortTitle":"Small agency WhatsApp and Slack"},{"ref":"MK-62417","title":"Long-form student documents and presentations bought for AI training","headline":"Long-form research papers, unpublished theses and presentations bought from student authors for AI training","description":"Long-form documents (research papers and unpublished theses across many domains) and presentations (finance, consulting, startups and more), bought from the student authors under agreements that let the owner use them for AI training. Content is screened for whether it is already public and anonymised. About 20,000 documents now, scalable to about 100,000, in English, Spanish, Italian and French.","category":"Text & documents","sector":"Education","subsector":"Student-authored academic and business documents","dataTypes":["Documents","Text"],"useCases":["Pre-training","Domain fine-tuning","SFT / instruction tuning"],"origin":"Licensed from third parties","regions":["North America","Europe"],"languages":["English","Spanish","Italian","French"],"sizeBand":"","count":"~20K documents","volume":"~20K documents, can scale to ~100K","years":"2009–2026","yearsOfHistory":"17","growing":"Yes","systems":[],"contents":["Research papers","Unpublished theses","Business and consulting presentations","Finance presentations","Startup decks"],"formats":["PDF","DOCX","PPT"],"delivery":"","labelled":"Partly","labelTypes":[],"personalData":"Low","deidentification":"Done by owner","rights":"Third-party licensed","sample":"Available","readiness":"Ready now","stage":"Summary available","scale":"","employees":"","industry":"Student-authored academic and business documents","tools":[],"span":"17 years","volumeHead":"20K documents","sizeTB":null,"fromYear":2009,"toYear":2026,"yearsNum":17.0,"tags":["theses","research papers","presentations","long-form documents","academic","student authored","multilingual"],"shortTitle":"Student documents and decks"},{"ref":"MK-63604","title":"Salesforce implementation history across 80+ client companies","headline":"Salesforce implementation record across 80+ client companies: requirements, workflows, automations, integrations and walkthrough videos","description":"A Salesforce consultancy's record of real implementations across 80+ client companies, some with over $1B in revenue, built up over 40,000+ billable hours. It covers business requirements, workflow and system-architecture design, automations, integrations, build decisions, task-level implementation details and technical problem-solving, plus recorded project discussions and walkthroughs. Org metadata (object and field schemas, configurations, flows, Apex code, components, integration structures) can also be compiled. Client details would be anonymised or redacted.","category":"Expert workflows","sector":"Professional services","subsector":"Salesforce consulting and implementation","dataTypes":["Documents","Code","Video","Tabular / structured"],"useCases":["Enterprise workflow agents","Coding models","Evaluation / benchmarks","SFT / instruction tuning","Computer-use & tool-use agents"],"origin":"First-party company records","regions":["North America","Europe"],"languages":["English"],"sizeBand":"","count":"40,000+ records","volume":"40,000+ records from 40,000+ billable hours; one folder per client org","years":"2022–2026","yearsOfHistory":"4","growing":"Yes","systems":["CRM","File storage"],"contents":["Business requirements","Workflow and process designs","System architectures","Automations and flows","Integrations","Build decisions","Task-level implementation records","Recorded walkthroughs","Apex code and org configuration","Object and field schemas"],"formats":["CSV","Code files","Video"],"delivery":"Download","labelled":"Yes","labelTypes":[],"personalData":"Moderate","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Ready now","stage":"Summary available","scale":"","employees":"","industry":"Salesforce consulting and implementation","tools":[["salesforce","Salesforce"],["files","File storage"]],"span":"4 years","volumeHead":"40K records","sizeTB":null,"fromYear":2022,"toYear":2026,"yearsNum":4.0,"tags":["salesforce","crm","consulting","enterprise workflows","automation","integrations","apex","implementation","business process"],"shortTitle":"Salesforce implementation history"},{"ref":"MK-66546","title":"Southern African commercial property developer: 15-year archive","headline":"Southern African commercial real estate developer and landlord: 15 years of email, files and accounting (~1 TB)","description":"About 1 TB covering 15 years from a Southern African commercial real estate developer and landlord working on mixed-use commercial projects with land owners, engineers, architects and contractors. Sources are email, cloud file storage and accounting records. The owner states no confidentiality agreements or court restrictions cover the loan, legal or contractor documents, and agrees to removal of tenant and contractor personal details.","category":"Company operations archive","sector":"Construction & real estate","subsector":"Commercial real estate development and leasing","dataTypes":["Email","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["Africa"],"languages":[],"sizeBand":"100 GB-1 TB","count":"~1 TB","volume":"~1 TB over 15 years","years":"15 years","yearsOfHistory":"15","growing":"","systems":["Email (Gmail, Outlook)","File storage (OneDrive)","Accounting (Sage Pastel)"],"contents":["Email with engineers, architects and contractors","Lease and tenant agreements","Loan and legal documents","Contractor documents","Accounting records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"Owner agrees to remove tenant and contractor personal details","rights":"Owner-held","sample":"","readiness":"Ready now","stage":"Summary available","scale":"","employees":"","industry":"Commercial real estate development and leasing","tools":[["gmail","Gmail"],["email","Outlook"],["files","OneDrive"],["accounting","Accounting"]],"span":"15 years","volumeHead":"1 TB","sizeTB":1.0,"fromYear":null,"toYear":null,"yearsNum":15.0,"tags":["commercial real estate","property development","landlord","leases","contractors","accounting","africa"],"shortTitle":"African commercial property files"},{"ref":"MK-70443","title":"Stereo egocentric video and handheld-gripper (UMI) data of manual work across 15+ industries","headline":"10K+ hours of stereo egocentric video and handheld-gripper demonstrations across 15+ industries","description":"Stereo first-person video (60 fps, 2x 1080p) and handheld-gripper (UMI-style) demonstration data of hard manual tasks across 15+ industries, recorded in the US, Philippines, Vietnam, Egypt, Pakistan and Turkey, capped at 5 hours per task and operator for diversity. 10K+ hours are available now and collection can scale to ~100K hours a month; a public test sample exists.","category":"Physical-world video","sector":"AI & data","subsector":"Egocentric / robot-learning data collection","dataTypes":["Video","Sensor / telemetry"],"useCases":["Robotics / embodied AI","World models","Vision & perception"],"origin":"Purpose-collected (consented capture)","regions":["North America","Southeast Asia","Middle East","South Asia"],"languages":[],"sizeBand":"","count":"10K+ hours","volume":"10K+ hours available; collection up to ~100K hours/month","years":"2026","yearsOfHistory":"","growing":"Yes","systems":[],"contents":["Stereo egocentric video","Gripper demonstration data","15+ industries","Hard manual tasks"],"formats":["MP4"],"delivery":"Download","labelled":"Partly","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held","sample":"Available","readiness":"Ready now","stage":"Summary available","scale":"","employees":"","industry":"Egocentric / robot-learning data collection","tools":[],"span":"","volumeHead":"10K hours","sizeTB":null,"fromYear":2026,"toYear":2026,"yearsNum":null,"tags":["egocentric","umi","robotics","manipulation","stereo video","physical ai","demonstrations"],"shortTitle":"Stereo egocentric work video"},{"ref":"MK-71362","title":"Indian AI consultancy: 100+ repositories plus email, chat, docs and wiki","headline":"AI consulting firm in India: 100+ code repositories from work for 50+ clients, plus Gmail, Slack, Drive and Notion, ~200 GB over 2.5 years","description":"About 2.5 years of data from an AI consulting firm in India that has served 50+ clients. Includes 100+ code repositories on GitHub, plus Gmail, Slack, Google Drive and Notion, about 200 GB in total. Data is ready now. Client code may carry client IP and rights questions that need confirming.","category":"Code & software","sector":"AI & data","subsector":"AI consulting","dataTypes":["Code","Email","Chat & messaging","Documents"],"useCases":["Coding models","Agentic coding","Enterprise workflow agents"],"origin":"First-party company records","regions":["South Asia"],"languages":[],"sizeBand":"100 GB-1 TB","count":"100+ repositories","volume":"~200 GB; 100+ code repositories; work for 50+ clients","years":"","yearsOfHistory":"2.5","growing":"","systems":["Gmail","Slack","Google Drive","Notion","GitHub"],"contents":["Code repositories","Email","Team chat","Drive documents","Notion pages"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Ready now","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"AI consulting","tools":[["slack","Slack"],["github","GitHub"],["gmail","Gmail"],["notion","Notion"],["googledrive","Google Drive"]],"span":"2.5 years","volumeHead":"100 repositories","sizeTB":0.2,"fromYear":null,"toYear":null,"yearsNum":2.5,"tags":["ai consulting","code","github","slack","notion","india"],"shortTitle":"Indian AI consultancy code archive"},{"ref":"MK-72241","title":"Open-finance startup IP: account aggregation, underwriting and transaction-categorisation code and docs","headline":"Defunct open-finance startup: code, design and docs for account aggregation, personal-finance, behavioural underwriting and a transaction-categorisation engine (~10 GB)","description":"Intellectual property of a closed open-finance venture: account-aggregation, personal-finance, behavioural-underwriting and transaction-categorisation code and backend services, design files, data-engineering docs, decks, explainer videos and internal mail. The categorisation engine was trained on transaction strings with personal data removed.","category":"Financial & transactional","sector":"Financial services & insurance","subsector":"Open finance / fintech","dataTypes":["Code","Documents","Video","Email"],"useCases":["Coding models","Domain fine-tuning"],"origin":"First-party company records","regions":["South Asia"],"languages":[],"sizeBand":"<10 GB","count":"","volume":"~10 GB","years":"~1 year","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Ready now","stage":"Summary available","scale":"","employees":"","industry":"Open finance / fintech","tools":[],"span":"1 year","volumeHead":"10 GB","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["fintech","open finance","underwriting","categorisation","code"],"shortTitle":"Open-finance startup codebase"},{"ref":"MK-78418","title":"3 years of Meta ad campaigns from a US marketing agency: 483 campaigns, $2.6M spend","headline":"3 years of Meta ad campaigns from a US marketing agency: 483 campaigns, $2.6M spend","description":"Meta (Facebook and Instagram) Ads Manager exports from a US marketing agency covering 483 campaigns across 4 ad accounts and about $2.6M in total spend, September 2023 to October 2026. Useful for ad-performance modelling, marketing agents and budget allocation.","category":"Consumer behaviour","sector":"Marketing, media & publishing","subsector":"Digital marketing agency","dataTypes":["Tabular / structured"],"useCases":["Forecasting & time series","Enterprise workflow agents","Domain fine-tuning"],"origin":"First-party company records","regions":["North America"],"languages":["English"],"sizeBand":"","count":"483 campaigns","volume":"483 campaigns; 4 ad accounts; ~$2.6M spend","years":"2023–2026","yearsOfHistory":"3","growing":"Yes","systems":["Analytics"],"contents":["Campaign settings and objectives","Spend by campaign","Impressions, clicks and conversions","4 ad accounts"],"formats":["CSV"],"delivery":"","labelled":"No","labelTypes":[],"personalData":"Low","deidentification":"Not needed","rights":"Owner-held, conditions to confirm","sample":"On request / under NDA","readiness":"Ready now","stage":"Summary available","scale":"","employees":"","industry":"Digital marketing agency","tools":[["analytics","Analytics & marketing"]],"span":"3 years","volumeHead":"483 campaigns","sizeTB":null,"fromYear":2023,"toYear":2026,"yearsNum":3.0,"tags":["meta ads","facebook ads","ad performance","marketing agency","campaign data"],"shortTitle":"Meta ad campaign history"},{"ref":"MK-81414","title":"Global survey-panel respondent profiles: ~30M records","headline":"Self-reported demographic, firmographic and behavioural profiles of survey respondents across 137 countries, with option to field new questions","description":"Self-reported demographic, firmographic and behavioural profiles of survey respondents, sourced through a survey exchange that aggregates 800+ supply partners (panels, apps and loyalty programmes) across 137 countries. ~30M records covering 2025-2026. Profiles refresh each time a respondent enters, and sessions are screened with a fraud/quality score. The owner can field new questions to the same population on request. Each respondent carries digital identifiers and/or a static respondent ID; no re-identification is permitted. Use is governed by respondent consent terms and a signed data exchange agreement, with model-training use scoped per deal and restrictions for EU/UK and some US states. Coverage skews to online survey takers and is thin in small markets.","category":"Consumer behaviour","sector":"Marketing, media & publishing","subsector":"Survey panel exchange","dataTypes":["Tabular / structured"],"useCases":["Domain fine-tuning","Forecasting & time series","Human preference / evaluation"],"origin":"Platform / user-generated","regions":["Global"],"languages":[],"sizeBand":"","count":"~30M respondent records","volume":"~30M respondent profile records; 137 countries; 800+ supply partners","years":"2025–2026","yearsOfHistory":"1","growing":"Yes","systems":[],"contents":["Demographic profiles","Firmographic profiles","Behavioural profiles","Respondent quality scores","Custom questions fielded on request"],"formats":["CSV","JSON","Parquet"],"delivery":"SFTP; S3/GCS; Snowflake","labelled":"Yes","labelTypes":[],"personalData":"Moderate","deidentification":"Pseudonymous respondent IDs; no re-identification permitted","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Ready now","stage":"Summary available","scale":"","employees":"","industry":"Survey panel exchange","tools":[],"span":"1 year","volumeHead":"30M respondent records","sizeTB":null,"fromYear":2025,"toYear":2026,"yearsNum":1.0,"tags":["survey panel","respondent profiles","demographics","firmographics","consumer behaviour","global"],"shortTitle":"Global survey respondent profiles"},{"ref":"MK-87719","title":"Video-game prediction-market startup: product analytics, support and sales","headline":"Video-game prediction-market startup: product analytics, support data and streamer sales","description":"Records of a venture-backed prediction-market platform for video games (~30K MAU, ~$40M GMV): product analytics database, customer-support data, outbound sales to onboard streamers, and company email over two years.","category":"Consumer behaviour","sector":"Consumer services","subsector":"Gaming prediction markets","dataTypes":["Tabular / structured","Email"],"useCases":["Domain fine-tuning","Enterprise workflow agents"],"origin":"First-party company records","regions":["North America"],"languages":["English"],"sizeBand":"1-10 TB","count":"~30K MAU","volume":"~1 TB","years":"2024–2026","yearsOfHistory":"2","growing":"","systems":["Database","Email"],"contents":["Product analytics","Customer support","Streamer outbound sales","Company email"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"Ready now","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Gaming prediction markets","tools":[["email","Email"],["database","Database"]],"span":"2 years","volumeHead":"30K MAU","sizeTB":1.0,"fromYear":2024,"toYear":2026,"yearsNum":2.0,"tags":["prediction markets","gaming","esports","analytics","customer support"],"shortTitle":"Game prediction-market startup"},{"ref":"MK-87734","title":"Drug-discovery chemistry lab: small-molecule synthesis task set","headline":"~2K small-molecule synthesis tasks from a drug-discovery chemistry lab, ~60% novel structures, including failure cases","description":"AI-trainable task dataset from a large chemistry lab doing small-molecule synthesis for drug discovery. About 2K compounds are ready now; around 60% are novel structures not found in PubChem, and the set includes failure cases. The lab produces many more compounds each month, though historically not in trainable form.","category":"Healthcare & life sciences","sector":"Healthcare & life sciences","subsector":"Small-molecule synthesis for drug discovery","dataTypes":[],"useCases":["Chemistry and drug-discovery models","Scientific reasoning tasks"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"~2K compounds","volume":"~2K compounds ready; more produced monthly","years":"","yearsOfHistory":"","growing":"Yes","systems":[],"contents":["Small-molecule synthesis tasks","Novel structures (~60% not in PubChem)","Failure cases"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Ready now","stage":"Summary available","scale":"Small team","employees":"200+","industry":"Small-molecule synthesis","tools":[],"span":"","volumeHead":"2K compounds","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["chemistry","drug discovery","small molecules","synthesis","novel compounds","failure cases","wet lab"],"shortTitle":"Novel small-molecule synthesis set"},{"ref":"MK-90009","title":"UK commercial real-estate loan performance","headline":"UK commercial real-estate loan performance: interest cover, valuations, payments and tenancy (~£1.5B of loans)","description":"Performance history for UK commercial real-estate loans totalling about £1.5B, covering interest cover ratios, valuations over time, whether payments were made on time, and tenant and rental information. Records are anonymised. Offered for models that predict how good a debt deal will be.","category":"Financial & transactional","sector":"Financial services & insurance","subsector":"Commercial real-estate lending","dataTypes":["Tabular / structured"],"useCases":["Forecasting & time series","Domain fine-tuning","Evaluation / benchmarks"],"origin":"First-party company records","regions":["Europe"],"languages":["English"],"sizeBand":"","count":"~£1.5B of loans","volume":"~£1.5B of loans","years":"2020–2026","yearsOfHistory":"6","growing":"","systems":[],"contents":["Interest cover ratios","Valuations over time","Payment history","Tenant information","Rental information"],"formats":[],"delivery":"","labelled":"Yes","labelTypes":[],"personalData":"Low","deidentification":"Done by owner","rights":"To be confirmed","sample":"On request / under NDA","readiness":"Ready now","stage":"Summary available","scale":"","employees":"","industry":"Commercial real-estate lending","tools":[],"span":"6 years","volumeHead":"~£1.5B of loans","sizeTB":null,"fromYear":2020,"toYear":2026,"yearsNum":6.0,"tags":["commercial real estate","loans","credit","debt","valuations","tenancy","uk","lending"],"shortTitle":"UK property loan performance"},{"ref":"MK-93138","title":"Industrial-site robotics simulation","headline":"Multi-agent robotics simulation of construction sites, warehouses, factories and water plants","description":"A multi-agent robotics/physics RL environment that simulates real-world construction sites, warehouses, factories and water plants, with workers as the action space, construction observations and a movement-based reward. It runs stochastically in the cloud behind a Python API at about 10,000 steps per second, with 1-second episodes, parallel runs, variable scenarios, ~20-minute setup and no special hardware. A baseline, demo and support are available; deliverables are data and deployment, tied to a vision behaviour model used for annotation. The seller states it holds the rights and that the environment has been validated with a frontier robotics lab.","category":"RL environment / benchmark","sector":"Construction & real estate","subsector":"Industrial-site robotics simulation","dataTypes":["RL environment / simulator","Video"],"useCases":["RL environments","Robotics / embodied AI","World models"],"origin":"Generated / simulated","regions":["North America"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"","systems":["Cloud simulator","API"],"contents":["Construction-site simulations","Warehouse simulations","Factory simulations","Water-plant simulations","Multi-agent worker actions","Movement-based reward","Baseline policy","Vision-model annotations"],"formats":["RL environment (Python API)"],"delivery":"API; Cloud-hosted","labelled":"Yes","labelTypes":["Vision behaviour model annotations"],"personalData":"None","deidentification":"","rights":"Owner-held","sample":"Available","readiness":"Ready now","stage":"Summary available","scale":"","employees":"","industry":"Industrial-site robotics simulation","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["robotics","simulation","construction","warehouse","factory","multi-agent","reinforcement learning","physics"],"shortTitle":"Industrial robotics simulation"},{"ref":"MK-11128","title":"Smart-shoe sensor data: gait, IMU, PPG and GPS at ~118 points/sec (~10 TB, 2 years)","headline":"Wearable smart-footwear data: IMU, PPG, foot temperature, GPS, heart rate/HRV/SpO2 and gait (stride, cadence, stance/swing, symmetry) across activity sessions","description":"Platform data from a smart-footwear product: inertial (accelerometer/gyroscope), PPG, foot temperature, GPS, heart rate, HRV, SpO2 and gait metrics across activity sessions. Pitched for humanoid and robotic locomotion.","category":"Industrial & sensor","sector":"Healthcare & life sciences","subsector":"Wearables","dataTypes":["Sensor / IMU","Tabular / structured"],"useCases":["Robotics & world models","Health models"],"origin":"Platform / user-generated","regions":["North America"],"languages":[],"sizeBand":"1-10 TB","count":"","volume":"~10 TB stated; ~118 sensor points per second; counts after audit","years":"~2 years","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Wearables","tools":[],"span":"2 years","volumeHead":"10 TB stated","sizeTB":10.0,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["wearable","gait","imu","locomotion","biometrics"],"shortTitle":"Smart-shoe sensor data"},{"ref":"MK-11571","title":"Indian residential rental agreements (~10K, 5 years)","headline":"~10K residential rental agreements (drafts and executed) from an online rental-agreement platform in India, over 5 years; may grow to ~100K","description":"Rental agreements prepared and registered through an online platform for landlords and tenants, in English, drafts and executed versions, over 5+ years. Volume may grow by combining several platforms.","category":"Text & documents","sector":"Legal","subsector":"Rental agreements","dataTypes":["Documents","Text"],"useCases":["Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["South Asia"],"languages":[],"sizeBand":"","count":"~10K agreements","volume":"~10K agreements (possibly ~100K)","years":"~5 years","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"Planned / on request","rights":"To be confirmed","sample":"Available","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Rental agreements","tools":[],"span":"5 years","volumeHead":"10K agreements","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["legal","contracts","rental","india"],"shortTitle":"Indian rental agreements"},{"ref":"MK-12165","title":"B2B lending RL environment: onboarding, underwriting and credit limits","headline":"B2B buy-now-pay-later lending simulation: agent onboards and underwrites small businesses and sets credit limits, rewarded against human decisions","description":"Web/REST environment simulating a B2B buy-now-pay-later lender: small-business onboarding, underwriting, transactions and disbursal. The agent scores applications and assigns limits; reward is within ±5% of the human decision, and human feedback is included.","category":"RL environment / benchmark","sector":"Financial services & insurance","subsector":"B2B lending","dataTypes":["RL environment / simulator"],"useCases":["RL environments","Enterprise workflow agents"],"origin":"Generated / simulated","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Within 2 weeks","stage":"Summary available","scale":"","employees":"","industry":"B2B lending","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["rl environment","lending","underwriting","credit"],"shortTitle":"B2B lending environment"},{"ref":"MK-12641","title":"Food and hospitality business: 10+ years of operating records","headline":"Food and hospitality business: 10-19 years of email, accounting, code, field operations and database records","description":"Operating data of a food and hospitality business in operation for 10 to 19 years. Includes email, accounting records, code repositories, field operations data and a MySQL database. The seller states it owns the data. Volumes are not stated.","category":"Company operations archive","sector":"Hospitality & food","subsector":"Food and hospitality business","dataTypes":["Email","Tabular / structured","Code"],"useCases":["Enterprise workflow agents","Coding models","Domain fine-tuning"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"10-19","growing":"","systems":["Email (Gmail)","Accounting (Xero)","Git (GitHub)","Field operations","Database (MySQL)"],"contents":["Email","Accounting records","Code repositories","Field operations data","Database"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"Mid-size company","employees":"51–200","industry":"Food and hospitality business","tools":[["github","GitHub"],["gmail","Gmail"],["database","Database"],["accounting","Accounting"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["food","hospitality","field operations","accounting","code","database"],"shortTitle":"Food business operating archive"},{"ref":"MK-12871","title":"Independent developer portfolio: SaaS, healthcare workflow, contract analysis and agent tooling code","headline":"Private code portfolio from one independent developer across SaaS operations, healthcare workflows, contract analysis and agent tooling, some with git history, tests and CI/CD","description":"Private projects from one independent developer, covering SaaS operations, healthcare workflows, contract analysis and agent tooling. Some projects keep their git history, tests, CI/CD configuration and architecture and deployment documents. The work is recent and partly AI-assisted. The seller offers a per-project inventory, without source, as a first step.","category":"Code & software","sector":"Software & SaaS","subsector":"Independent software development","dataTypes":["Code","Documents"],"useCases":["Coding models","Agentic coding"],"origin":"First-party company records","regions":["Europe"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"","systems":["Git"],"contents":["SaaS operations projects","Healthcare workflow projects","Contract analysis projects","Agent tooling","Tests and CI/CD","Architecture and deployment docs"],"formats":[],"delivery":"","labelled":"No","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Timing to confirm","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Independent software development","tools":[["git","Git"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["code","saas","healthcare workflows","contract analysis","agent tooling","ci/cd"],"shortTitle":"Solo developer project portfolio"},{"ref":"MK-14916","title":"Consumer app and marketplace: email and cloud files","headline":"Consumer app and marketplace with 1M+ customers: 5-9 years of email and cloud file storage","description":"Internal records of a consumer app and marketplace with over 1M customers, in operation for 5 to 9 years: business email and cloud file storage across two services. The seller states it owns the data. Volumes are not stated.","category":"Company operations archive","sector":"Software & internet","subsector":"Consumer app and marketplace","dataTypes":["Email","Documents"],"useCases":["Enterprise workflow agents","Retrieval / RAG","Domain fine-tuning"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"5-9","growing":"","systems":["Email (Gmail)","File storage (Google Drive, Dropbox)"],"contents":["Business email","Cloud documents and files"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Consumer app and marketplace","tools":[["gmail","Gmail"],["googledrive","Google Drive"],["dropbox","Dropbox"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["consumer app","marketplace","email","documents"],"shortTitle":"Consumer marketplace app files"},{"ref":"MK-16037","title":"Engineering firm: projects, bids and drawings","headline":"Brazilian engineering and construction firm: project histories, bids, estimates, site reports and drawings","description":"Records from a long-established engineering and construction firm in Brazil: engineering project histories, bids and estimates, site reports and engineering drawings going back many years. Volumes and formats have not yet been provided; the owner prefers a call before completing an inventory.","category":"Company operations archive","sector":"Construction & real estate","subsector":"Engineering and construction firm","dataTypes":["Documents","3D / CAD","Tabular / structured","Text"],"useCases":["Domain fine-tuning","Enterprise workflow agents","Vision & perception","Retrieval / RAG"],"origin":"First-party company records","regions":["Latin America"],"languages":[],"sizeBand":"","count":"","volume":"","years":"Multiple years","yearsOfHistory":"","growing":"","systems":["File storage","Design files"],"contents":["Engineering project histories","Bids","Cost estimates","Site reports","Engineering drawings"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Mid-size company","employees":"51–200","industry":"Engineering and construction firm","tools":[["design","Design files"],["files","File storage"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["engineering","construction","bids","estimates","site reports","drawings","brazil"],"shortTitle":"Engineering firm bids and drawings"},{"ref":"MK-16252","title":"E-commerce accounting-sync app: 9 years of support tickets, accounting records and code","headline":"Small North American software company behind an e-commerce accounting integration: email, accounting records, support tickets and GitHub code over 9 years","description":"Data from a small North American software company that runs an integration syncing e-commerce store data with accounting software (QuickBooks and Xero). Includes email from 5 inboxes, QuickBooks and Xero records, customer support tickets from Zendesk and Intercom, and GitHub code, about 2 GB over 9 years. The customer base is in the thousands. Export is estimated at 1-2 weeks.","category":"Company operations archive","sector":"Software & SaaS","subsector":"E-commerce accounting integration","dataTypes":["Email","Tabular / structured","Text","Code"],"useCases":["Enterprise workflow agents","Customer support agents","Coding models"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"<10 GB","count":"~2 GB","volume":"~2 GB; 5 email inboxes; support tickets; accounting records; code","years":"","yearsOfHistory":"9","growing":"","systems":["Gmail","QuickBooks","Xero","Zendesk","Intercom","GitHub"],"contents":["Email","Accounting records","Customer support tickets","Source code"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"Export 1-2 weeks","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"E-commerce accounting integration","tools":[["github","GitHub"],["zendesk","Zendesk"],["intercom","Intercom"],["gmail","Gmail"],["quickbooks","QuickBooks"]],"span":"9 years","volumeHead":"2 GB","sizeTB":0.002,"fromYear":null,"toYear":null,"yearsNum":9.0,"tags":["saas","e-commerce","accounting","integration","support tickets","code"],"shortTitle":"E-commerce accounting app archive"},{"ref":"MK-16394","title":"Long-horizon cybersecurity agent environments","headline":"Resettable long-horizon cybersecurity environments built on real vulnerabilities, deterministic grading","description":"Resettable cyber environments for training and evaluating security agents on long-horizon tasks based on real vulnerabilities: supply-chain attacks, multi-stage application exploitation, protocol vulnerabilities and patch bypasses. Grading is deterministic with no LLM judges. Built in-house; number of environments and delivery format not yet stated.","category":"RL environment / benchmark","sector":"Software & internet","subsector":"Cybersecurity","dataTypes":["RL environment / simulator","Code"],"useCases":["RL environments","Evaluation / benchmarks","Computer-use & tool-use agents","Coding models"],"origin":"Expert-created","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":["Supply-chain attack scenarios","Multi-stage application exploitation","Protocol vulnerability tasks","Patch-bypass tasks","Deterministic graders","Resettable environment state"],"formats":["RL / evaluation environments"],"delivery":"","labelled":"Yes","labelTypes":["Deterministic grading"],"personalData":"None","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Cybersecurity","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["cybersecurity","security agents","vulnerabilities","exploitation","reinforcement learning","long-horizon","evals"],"shortTitle":"Cybersecurity agent environments"},{"ref":"MK-16420","title":"Private-equity firm deal and communications dataset","headline":"Private-equity firm dataset (~1 TB): deal data, internal chat, email, meeting notes and versioned slide decks","description":"An internal dataset from a private-equity firm, offered through an intermediary on the firm's behalf: deal data, internal chat messages, emails, meeting notes and slide decks including their version histories, about 1 TB in total. The intermediary also captures long-horizon work traces from traditional businesses and enterprises and is open to processing or refining datasets with partners. The dataset is being prepared for sale.","category":"Company operations archive","sector":"Financial services & insurance","subsector":"Private equity firm","dataTypes":["Email","Chat & messaging","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Pre-training","Retrieval / RAG","SFT / instruction tuning"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"1-10 TB","count":"~1 TB","volume":"~1 TB","years":"","yearsOfHistory":"","growing":"","systems":["Email","Team chat","File storage"],"contents":["Deal data","Internal chat messages","Emails","Meeting notes","Slide decks with version histories"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Via authorised intermediary","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Private equity firm","tools":[["chat","Team chat"],["email","Email"],["files","File storage"]],"span":"","volumeHead":"1 TB","sizeTB":1.0,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["private equity","deal data","email","slack","meeting notes","slide decks","version history","finance"],"shortTitle":"Private-equity deal records"},{"ref":"MK-17373","title":"Web3 growth platform: 6 years of CRM, email (30 inboxes), GitHub, SOPs and finance records (~1 TB)","headline":"Web3 growth and analytics platform: 5 years of HubSpot CRM, 30 mailboxes, GitHub, Drive with SOPs, ERP and invoices","description":"Operating records of a web3 growth platform (analytics, automation, engagement): CRM, shared drive with SOPs, ERP and invoices, 30 email inboxes and code repositories.","category":"Company operations archive","sector":"Software & SaaS","subsector":"Web3 / marketing tech","dataTypes":["Code","Email","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Coding models"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"1-10 TB","count":"","volume":"~1 TB","years":"~6 years","yearsOfHistory":"","growing":"","systems":["HubSpot","Google Drive","ERP","GitHub"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Medium","deidentification":"","rights":"Owner-held","sample":"","readiness":"Export about a month","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Web3 / marketing tech","tools":[["github","GitHub"],["hubspot","HubSpot"],["googledrive","Google Drive"],["erp","ERP"]],"span":"6 years","volumeHead":"1 TB","sizeTB":1.0,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["web3","crm","github","sops"],"shortTitle":"Web3 growth platform archive"},{"ref":"MK-17858","title":"Indian D2C intimate-apparel brand: 1.5-year company data","headline":"Indian direct-to-consumer intimate-apparel brand: ~10 GB of company data including customer orders","description":"Company data from a small Indian direct-to-consumer intimate-apparel brand, about 10 GB covering 1.5 years, including D2C customer orders. The founder wants to sell and has offered to help extract the data. Systems and record counts are not yet stated.","category":"Company operations archive","sector":"Retail & e-commerce","subsector":"D2C intimate-apparel brand","dataTypes":["Tabular / structured"],"useCases":["Enterprise workflow agents","Forecasting & time series"],"origin":"First-party company records","regions":["South Asia"],"languages":[],"sizeBand":"10-100 GB","count":"~10 GB","volume":"~10 GB; 1.5 years; customer orders","years":"1.5 years","yearsOfHistory":"1.5","growing":"","systems":[],"contents":["D2C customer orders"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"D2C intimate-apparel brand","tools":[],"span":"1.5 years","volumeHead":"10 GB","sizeTB":0.01,"fromYear":null,"toYear":null,"yearsNum":1.5,"tags":["d2c","e-commerce","apparel","customer orders","india"],"shortTitle":"Indian D2C apparel brand records"},{"ref":"MK-18205","title":"Insurance agency: 10-19 years of email, chat, CRM, call recordings and accounting","headline":"Insurance agency with tens of thousands of customers: email, team chat, CRM, accounting, support, call recordings, files, projects and contracts","description":"Operating data from an insurance agency with 10-19 years of history and a customer base in the tens of thousands. The seller lists Outlook and Gmail, Slack, CRM, accounting, support, call recordings, files, projects and contracts. Volumes have not been provided yet. Insurance records carry sensitive customer personal data.","category":"Company operations archive","sector":"Financial services & insurance","subsector":"Insurance agency","dataTypes":["Email","Chat & messaging","Tabular / structured","Audio / speech","Documents"],"useCases":["Enterprise workflow agents","Speech & audio models","Customer support agents","Domain fine-tuning"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"10-19","growing":"","systems":["Email (Outlook","Gmail)","Slack","CRM","Accounting","Support desk","Call recordings","File storage","Project management"],"contents":["Email","Team chat","CRM records","Accounting records","Support conversations","Call recordings","Files","Projects","Contracts"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Insurance agency","tools":[["slack","Slack"],["gmail","Gmail"],["email","Outlook"],["tickets","Tickets"],["crm","CRM"],["calls","Calls"],["accounting","Accounting"],["files","File storage"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["insurance","agency","call recordings","crm","email","customer service"],"shortTitle":"Insurance agency call archive"},{"ref":"MK-18301","title":"Small business: email, shared files and contracts","headline":"Small business with thousands of customers offering 5-9 years of email, shared files and contracts","description":"Operating data of a small business (sector not specified) with thousands of customers, covering 5-9 years: email, shared files and contracts. Volumes not yet stated.","category":"Company operations archive","sector":"Other","subsector":"","dataTypes":["Email","Documents"],"useCases":["Enterprise workflow agents","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"Thousands of customers","years":"","yearsOfHistory":"5-9","growing":"","systems":["Email","File storage"],"contents":["Email","Shared files","Contracts"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Other","tools":[["email","Email"],["files","File storage"]],"span":"","volumeHead":"Thousands of customers","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["small business","email","contracts","shared files"],"shortTitle":"Small-firm email and contracts"},{"ref":"MK-19597","title":"Fleet driving data from dense roads: vision, behaviour and sensors","headline":"Daily driving data from a fleet of 50+ cars on dense roads: camera video, driver behaviour and vehicle sensor data","description":"Driving data collected daily from a fleet of 50+ cars on dense, high-traffic roads, combining vision (camera) data, driver-behaviour data and vehicle sensor data. The owner describes it as compliance friendly. Footage includes street scenes with pedestrians and number plates. Location, total volume and rights terms are not yet stated.","category":"Physical-world video","sector":"Automotive & transport","subsector":"Vehicle fleet driving data","dataTypes":["Video","Sensor / telemetry"],"useCases":["Vision & perception","Autonomous driving & robotics"],"origin":"","regions":["Not stated"],"languages":[],"sizeBand":"","count":"50+ cars collecting daily","volume":"50+ cars collecting daily","years":"","yearsOfHistory":"","growing":"Yes","systems":[],"contents":["Vehicle camera footage","Driver behaviour data","Vehicle sensor data"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Vehicle fleet driving data","tools":[],"span":"","volumeHead":"50 cars collecting daily","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["driving data","dashcam","dense traffic","driver behaviour","sensors","autonomous driving"],"shortTitle":"Dense-traffic fleet driving data"},{"ref":"MK-20070","title":"Roofing contractors: operating data (multiple companies)","headline":"Pipeline of ~150 US roofing contractors ($5M-$125M revenue) offering operating data; first inventory received","description":"Business data from roofing contractors represented by a business adviser who works with about 150 roofing companies ranging from roughly $5M to $125M in revenue. The first contractor has returned a detailed data inventory (listed separately) and more contractors are being approached. Contents per company will vary; nothing beyond the first inventory is described yet.","category":"Company operations archive","sector":"Construction & real estate","subsector":"Roofing contractors (multi-company pipeline)","dataTypes":["Tabular / structured","Documents"],"useCases":["Enterprise workflow agents","Domain fine-tuning"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"","count":"~150 contractors in pipeline","volume":"~150 contractors in pipeline","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":["Roofing contractor business records","Per-company data inventories (first one received)"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Via authorised intermediary","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Roofing contractors","tools":[],"span":"","volumeHead":"150 contractors in pipeline","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["roofing","contractors","construction","home services","multi-company","pipeline"],"shortTitle":"Roofing contractors data"},{"ref":"MK-20713","title":"Automotive component plant: egocentric smart-glasses video of machining, welding and QC (collected to order)","headline":"Head-mounted video of CNC setup, welding, shaft handling and quality control at an automotive component manufacturer, with frame-accurate step labels","description":"A capture team offers to record egocentric video at an automotive component manufacturer using head-mounted smart glasses, plus multi-angle footage. Tasks include CNC machine setup, welding, transmission shaft handling and quality control, with frame-accurate step labels. The plan is a pilot (10-15 hours stated, up to ~50 hours) on 2-3 stations, then scaling across the manufacturer's two plants. Footage has not been captured yet; the pilot can start within two weeks. Site management is aligned and worker consent is to be collected under India's data protection law.","category":"Physical-world video","sector":"Manufacturing","subsector":"Automotive component manufacturing","dataTypes":["Video","Labels"],"useCases":["Robotics & embodied AI","Vision & perception","Video understanding"],"origin":"Purpose-collected (consented capture)","regions":["South Asia"],"languages":[],"sizeBand":"","count":"10-50 hours (pilot)","volume":"Pilot of 10-15 hours stated, with ~50 hours mentioned; 2-3 stations initially; scale-up across two plants","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":["CNC machine setup","Welding","Transmission shaft handling","Quality control","Frame-accurate step labels"],"formats":["Egocentric and multi-angle video with labels"],"delivery":"","labelled":"Yes","labelTypes":["Frame-accurate step labels"],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"Not available","readiness":"Collected to order","stage":"Collected to order","scale":"","employees":"","industry":"Automotive component manufacturing","tools":[],"span":"","volumeHead":"10-50 hours","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["egocentric video","smart glasses","manufacturing","automotive","welding","cnc","quality control","step labels"],"shortTitle":"Smart-glasses machining video"},{"ref":"MK-21451","title":"Closed consumer startup: 18 repos across mobile, web, backend and a 3D app","headline":"Closed memorial-services startup: 18 repos with 319 PRs (Flutter mobile, TypeScript web/admin, Python backend, Unity 3D)","description":"Code-only archive of a closed consumer startup: 7 core repos (backend, Flutter mobile app, admin, web app, panorama server, Unity 3D app) and 18 repos with 319 pull requests overall, in TypeScript/JavaScript, Dart, Python and C#. Customer records, inboxes and documents excluded.","category":"Code & software","sector":"Software & internet","subsector":"Consumer app","dataTypes":["Code"],"useCases":["Coding models"],"origin":"First-party company records","regions":["Southeast Asia"],"languages":[],"sizeBand":"<10 GB","count":"18 repos; 319 PRs","volume":"~0.2 GB incl. Git history; 18 repos; 319 PRs","years":"2023–2024","yearsOfHistory":"1","growing":"","systems":["Git"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Consumer app","tools":[["git","Git"]],"span":"1 year","volumeHead":"18 repos","sizeTB":null,"fromYear":2023,"toYear":2024,"yearsNum":1.0,"tags":["flutter","typescript","unity","repos"],"shortTitle":"Closed consumer startup repos"},{"ref":"MK-21646","title":"Video segment-selection (trim) events","headline":"~3M real user decisions on which segment of an online video to clip, with timestamps and session metadata","description":"Roughly 3 million trim events from a consumer video-clipping tool, each capturing a real person's choice of which segment of an online video to cut: video title and duration, chosen start and end timestamps, clip duration and output format, coarse country/location, plan tier (free vs paid), event and session timestamps, and device/session metadata with pseudonymous user and session IDs. Collected by the seller's own platform over the past few years and available as a historical licence or an ongoing feed.","category":"Consumer behaviour","sector":"Software & internet","subsector":"Consumer video-clipping tool","dataTypes":["Tabular / structured","Text"],"useCases":["Preference / RLHF","Vision & perception","Evaluation / benchmarks","SFT / instruction tuning"],"origin":"Platform / user-generated","regions":["Global"],"languages":[],"sizeBand":"","count":"~3M trim events","volume":"~3M trim events","years":"2026","yearsOfHistory":"","growing":"Yes","systems":["Analytics","Application database"],"contents":["Video title and duration","User-selected start/end timestamps","Clip duration","Output format","Country/location metadata","Free vs paid plan tier","Event and session timestamps","Device/session metadata","Pseudonymous user and session IDs"],"formats":[],"delivery":"Historical licence; Ongoing feed","labelled":"Partly","labelTypes":["Human segment selections (start/end timestamps)"],"personalData":"Moderate","deidentification":"Planned / on request","rights":"Owner-held","sample":"Available","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Consumer video-clipping tool","tools":[["database","Database"],["analytics","Analytics & marketing"]],"span":"","volumeHead":"3M trim events","sizeTB":null,"fromYear":null,"toYear":2026,"yearsNum":null,"tags":["video","clipping","highlights","human preference","timestamps","user behaviour","segment selection"],"shortTitle":"Video trim events"},{"ref":"MK-22212","title":"Small real estate firm: 10-19 year operations archive","headline":"Small real estate firm: 10-19 years of email, accounting, calls, files, orders, contracts and HR","description":"Company data from a small real estate business with thousands of customers and 10-19 years of history. Systems listed: email, accounting, calls, file storage, orders/POS, contracts and HR records. Volumes are not yet stated.","category":"Company operations archive","sector":"Construction & real estate","subsector":"Real estate firm","dataTypes":["Email","Audio / speech","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Speech & audio models"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"10-19 years","yearsOfHistory":"10-19","growing":"","systems":["Email","Accounting","Telephony","File storage","POS","Contracts","HR"],"contents":["Email","Accounting records","Calls","Files","Orders/POS","Contracts","HR records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Real estate firm","tools":[["email","Email"],["calls","Calls"],["accounting","Accounting"],["shop","Commerce"],["hr","HR & payroll"],["files","File storage"]],"span":"19 years","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["real estate","contracts","calls","accounting","long history"],"shortTitle":"Real estate firm 10-19yr records"},{"ref":"MK-22833","title":"US multifamily real estate platform: 15-year archive incl. construction costs, leasing and maintenance","headline":"~50 TB from a large vertically integrated US multifamily platform: email, Teams, CRM, ERP, CAD and construction costs, unit pricing and leasing, resident behaviour and maintenance","description":"Fifteen years of data from a large, vertically integrated multifamily real estate platform in the US. Covers email and Teams, CRM, ERP, support, code, SharePoint documents, field operations, CAD, takeoffs and construction costs, unit pricing and leasing, resident behaviour, maintenance and HR. Total size is stated as ~50 TB across email, SharePoint and Azure. The submitter is the technology and data lead, so company sign-off is not yet confirmed. Export is estimated at 1-2 weeks.","category":"Company operations archive","sector":"Construction & real estate","subsector":"Multifamily real estate (development, leasing and management)","dataTypes":["Email","Chat & messaging","Documents","Tabular / structured","Code","CAD / drawings"],"useCases":["Enterprise workflow agents","Forecasting & time series","Pricing & demand models","Domain fine-tuning"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"10-100 TB","count":"~50 TB","volume":"~50 TB across email, SharePoint and Azure","years":"","yearsOfHistory":"15","growing":"","systems":["Outlook","Teams","CRM","ERP","Support desk","Code repository","SharePoint","Azure","Field operations","CAD"],"contents":["Email and Teams","CRM and ERP records","Support records","Code","SharePoint documents","Field operations","CAD, takeoffs and construction costs","Unit pricing and leasing","Resident behaviour","Maintenance","HR"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Export 1-2 weeks","stage":"Summary available","scale":"Small team","employees":"200+","industry":"Multifamily real estate","tools":[["email","Outlook"],["files","SharePoint"],["git","Git"],["crm","CRM"],["erp","ERP"],["design","Design files"]],"span":"15 years","volumeHead":"50 TB","sizeTB":50.0,"fromYear":null,"toYear":null,"yearsNum":15.0,"tags":["multifamily","real estate","leasing","pricing","construction costs","cad","maintenance","property management"],"shortTitle":"US multifamily operator archive"},{"ref":"MK-23328","title":"European fabrication company in insolvency: 30 years of technical drawings on intact servers","headline":"Three decades of engineering data from a large European fabrication company in insolvency, with intact servers and a large in-house technical drawing team","description":"Data from a European fabrication company with roughly 30 years of operations, now in insolvency. A large share of its staff worked only on technical drawings, and the company servers are reported intact and untampered. The source is in contact with the insolvency administrator to start the process. File counts, formats and volumes have not been provided yet.","category":"Company operations archive","sector":"Industrial & engineering","subsector":"Fabrication","dataTypes":["CAD / drawings","Documents"],"useCases":["Engineering & CAD models","Vision & perception","Domain fine-tuning"],"origin":"First-party company records","regions":["Europe"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"~30","growing":"No","systems":["File servers"],"contents":["Technical drawings","Company server contents"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Timing to confirm","stage":"Summary available","scale":"Large organisation","employees":"200+","industry":"Fabrication","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["fabrication","engineering drawings","cad","insolvency","manufacturing","europe"],"shortTitle":"Fabricator engineering drawings"},{"ref":"MK-24537","title":"Enterprise software process videos and SOPs (SAP, Salesforce and 100s of apps)","headline":"~25 TB of enterprise-software process videos with matching SOPs (SAP, Salesforce and hundreds more), 5 years","description":"Screen-recorded videos showing how business processes are executed in enterprise software such as SAP, Salesforce, Autodesk and Infor, covering end-to-end flows like procure-to-pay and hire-to-retire across thousands of companies. Each process comes with step-by-step SOPs and markdown documents describing the same flow, plus long-horizon material on how these processes are transformed. A separate database holds process flows across hundreds of software applications. The owner is an India-based platform company; much of the material was recorded by or for its customers, so rights to customer content are to be confirmed.","category":"Agent traces","sector":"Software & internet","subsector":"Process video and SOP documentation platform","dataTypes":["Video","Screen recordings / UI traces","Documents","Text"],"useCases":["Computer-use & tool-use agents","Enterprise workflow agents","SFT / instruction tuning","Evaluation / benchmarks","Vision & perception"],"origin":"Platform / user-generated","regions":["Global"],"languages":[],"sizeBand":"10-100 TB","count":"~25 TB; thousands of companies","volume":"~25 TB","years":"2026","yearsOfHistory":"5","growing":"","systems":["ERP","CRM","Design files","File storage","Database"],"contents":["Screen-recorded process execution videos","SOP documents paired with each video","Markdown process descriptions","Procure-to-pay flows","Hire-to-retire flows","SAP, Salesforce, Autodesk and Infor processes","Process-flow database covering hundreds of applications","Long-horizon process-transformation material"],"formats":["Video","Markdown","SOP documents"],"delivery":"","labelled":"Partly","labelTypes":["Step-by-step SOPs paired with videos"],"personalData":"Moderate","deidentification":"Done by owner","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Within 2 weeks","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Process video and SOP documentation platform","tools":[["salesforce","Salesforce"],["sap","SAP"],["database","Database"],["design","Design files"],["files","File storage"]],"span":"5 years","volumeHead":"25 TB","sizeTB":25.0,"fromYear":null,"toYear":2026,"yearsNum":5.0,"tags":["screen recordings","sop","process videos","sap","salesforce","procure-to-pay","hire-to-retire","enterprise workflows","computer use","erp"],"shortTitle":"Enterprise software SOP videos"},{"ref":"MK-25078","title":"Resale e-commerce: sale prices, bid/ask, listings and inventory sell-through","headline":"Resale e-commerce: timestamped sale prices, bid/ask, listings and inventory sell-through","description":"Data from a platform for online resellers of physical goods: timestamped sale prices and bid/ask spreads per product across major resale marketplaces, listing metadata (type, titles, descriptions), and anonymised inventory statistics such as inventory value over time, sell-through, margins and unit cost. Products are keyed by UPC. Inventory data is the owner's; some price and listing data is gathered from marketplaces.","category":"Consumer behaviour","sector":"Retail & e-commerce","subsector":"Reseller inventory platform","dataTypes":["Tabular / structured","Text"],"useCases":["Forecasting & time series","Enterprise workflow agents","Domain fine-tuning"],"origin":"Mixed","regions":["North America"],"languages":["English"],"sizeBand":"100 GB-1 TB","count":"Hundreds of GB","volume":"Hundreds of GB","years":"2025–2026","yearsOfHistory":"1","growing":"Yes","systems":[],"contents":["Sale prices per marketplace","Bid/ask time series","Listing metadata","Inventory sell-through","Margins and unit cost"],"formats":["CSV"],"delivery":"Cloud bucket (S3)","labelled":"Yes","labelTypes":[],"personalData":"None","deidentification":"Not needed","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Within 2 weeks","stage":"Summary available","scale":"","employees":"","industry":"Reseller inventory platform","tools":[],"span":"1 year","volumeHead":"Hundreds of GB","sizeTB":null,"fromYear":2025,"toYear":2026,"yearsNum":1.0,"tags":["resale","ecommerce","pricing","bid ask","inventory","sneakers","marketplaces"],"shortTitle":"Resale marketplace pricing"},{"ref":"MK-25712","title":"Legal-education company: 10 years of courses, sales calls, CRM, documents and HR records (~50 TB)","headline":"Established legal-education company: ~50 TB over 10 years incl. tens of thousands of sales calls with transcripts, CRM, course records and documents","description":"Online legal-education business teaching law, accounting and finance professionals. Course records, tens of thousands of recorded sales calls with transcripts, CRM support and sales history, shared documents, HR and invoicing records, advertising data and staff-activity records. The owner can also supply a community of law and finance professionals to validate model outputs.","category":"Company operations archive","sector":"Education","subsector":"Legal and professional education","dataTypes":["Audio / speech","Text","Email","Documents","Tabular / structured"],"useCases":["Domain fine-tuning","Enterprise workflow agents","Speech & audio models"],"origin":"First-party company records","regions":["South Asia"],"languages":[],"sizeBand":"10-100 TB","count":"Tens of thousands of sales calls","volume":"~50 TB; tens of thousands of sales calls; ~500 mailboxes","years":"2016–2026","yearsOfHistory":"10","growing":"","systems":["CRM","File storage","HR","Accounting","Email"],"contents":["Course records","Recorded sales calls with transcripts","CRM support and sales","Documents","HR and invoices","Ad data"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"On request / under NDA","readiness":"On request","stage":"Summary available","scale":"Large organisation","employees":"200+","industry":"Legal and professional education","tools":[["email","Email"],["crm","CRM"],["accounting","Accounting"],["hr","HR & payroll"],["files","File storage"]],"span":"10 years","volumeHead":"Tens of thousands of sales calls","sizeTB":50.0,"fromYear":2016,"toYear":2026,"yearsNum":10.0,"tags":["legal","education","sales calls","transcripts","crm","india"],"shortTitle":"Legal-education company archive"},{"ref":"MK-27101","title":"Small marketing agency: 10+ years of internal systems","headline":"Small marketing agency: 10-19 years of email, chat, CRM, accounting, code, database and e-signed documents","description":"Internal systems of a small marketing agency with under 100 customers and 10 to 19 years in operation: email, team chat, CRM, accounting, Git code repositories, cloud documents, a PostgreSQL database and e-signature documents. The seller says the data is most likely company-owned. Volumes are not stated.","category":"Company operations archive","sector":"Marketing, media & publishing","subsector":"Marketing agency","dataTypes":["Email","Chat & messaging","Code","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Coding models","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"10-19","growing":"","systems":["Email (Gmail)","Team chat (Slack)","CRM (HubSpot)","Accounting (QuickBooks)","Git (GitHub)","File storage (Google Drive)","Database (PostgreSQL)","E-signature (PandaDoc, Acrobat Sign)"],"contents":["Email","Team chat","CRM records","Accounting","Code repositories","Documents","Database","E-signed documents"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Marketing agency","tools":[["slack","Slack"],["github","GitHub"],["hubspot","HubSpot"],["gmail","Gmail"],["googledrive","Google Drive"],["quickbooks","QuickBooks"],["database","Database"],["sign","E-signature"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["marketing agency","hubspot","slack","code","crm","agency operations"],"shortTitle":"Marketing agency systems archive"},{"ref":"MK-28433","title":"Small digital marketing agency: 5-9 year company archive","headline":"Small digital marketing agency: 5-9 years of email, chat, accounting, code, call recordings, Drive files and contracts","description":"Company data from a small marketing agency with under 100 customers and 5-9 years of history. Systems listed: email, team chat and messaging, accounting, code repositories, video-call recordings, file storage and e-signed contracts. Volumes are not yet stated.","category":"Company operations archive","sector":"Marketing, media & publishing","subsector":"Digital marketing agency","dataTypes":["Email","Chat & messaging","Code","Video","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Coding models","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"5-9 years","yearsOfHistory":"5-9","growing":"","systems":["Email (Gmail)","Team chat (Slack)","WhatsApp","Accounting (Xero)","GitHub","Video calls (Zoom, Meet, Teams)","File storage (Google Drive)","E-signature"],"contents":["Email","Slack and WhatsApp chat","Xero accounting","GitHub code","Zoom/Meet/Teams recordings","Drive files","Contracts"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Digital marketing agency","tools":[["slack","Slack"],["whatsapp","WhatsApp"],["github","GitHub"],["gmail","Gmail"],["googledrive","Google Drive"],["zoom","Zoom"],["accounting","Accounting"],["hr","HR & payroll"],["sign","E-signature"]],"span":"9 years","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["marketing agency","client work","code","call recordings","accounting"],"shortTitle":"Digital marketing agency archive"},{"ref":"MK-29757","title":"Large construction company: 10-19 year operations archive","headline":"Large construction company: 10-19 years of email, chat, CRM, accounting, recordings, projects, field ops, CAD and contracts","description":"Company data from a large construction business with 10-19 years of history. Systems listed: email, team chat and messaging, CRM, accounting, recordings, files, project records, orders, field operations, CAD, a database, contracts and HR records. Volumes are not yet stated and company size is still to be verified.","category":"Company operations archive","sector":"Construction & real estate","subsector":"Construction contractor","dataTypes":["Email","Chat & messaging","Documents","CAD / 3D","Audio / speech","Tabular / structured"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Vision & perception","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"10-19 years","yearsOfHistory":"10-19","growing":"","systems":["Email (Gmail)","Team chat (Slack, Teams)","WhatsApp","CRM (Salesforce)","Accounting (QuickBooks)","Recordings","File storage","Project management","Field operations","CAD","Database","Contracts","HR"],"contents":["Email","Slack, Teams and WhatsApp chat","Salesforce CRM","QuickBooks accounting","Recordings","Project files","Orders","Field operations records","CAD drawings","Database","Contracts","HR records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"Large organisation","employees":"200+","industry":"Construction contractor","tools":[["slack","Slack"],["whatsapp","WhatsApp"],["salesforce","Salesforce"],["gmail","Gmail"],["quickbooks","QuickBooks"],["tickets","Tickets"],["database","Database"],["design","Design files"],["hr","HR & payroll"],["files","File storage"]],"span":"19 years","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["construction","field operations","cad","projects","crm","long history"],"shortTitle":"Construction firm 10-19 year ops"},{"ref":"MK-30797","title":"Legal expert demonstrations and contract-negotiation trajectories (to order, via law-firm network)","headline":"Legal expert demonstrations and contract negotiation/redlining trajectories, produced to order","description":"A legal AI company with its own legal environments and workflows and a network of law firms and lawyers offers to produce legal training data to a buyer's specification: expert demonstrations, annotated legal work, contract negotiation and redlining trajectories, evaluations, RL environments or custom datasets. No data has been produced yet; supply depends on participating law firms agreeing to license.","category":"Text & documents","sector":"Legal","subsector":"Legal AI / law-firm network","dataTypes":["Text","Documents","Agent trajectories"],"useCases":["SFT / instruction tuning","RL environments","Evaluation / benchmarks","Enterprise workflow agents"],"origin":"Expert-created","regions":["Europe","North America"],"languages":[],"sizeBand":"","count":"","volume":"To order","years":"To order","yearsOfHistory":"","growing":"","systems":[],"contents":["Expert legal demonstrations","Annotated legal work","Contract negotiation trajectories","Redlining trajectories","Legal evaluations","Legal RL environments"],"formats":[],"delivery":"","labelled":"Yes","labelTypes":["Expert annotations"],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"Not available","readiness":"Collected to order","stage":"Collected to order","scale":"","employees":"","industry":"Legal AI / law-firm network","tools":[],"span":"","volumeHead":"To order","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["legal","contracts","redlining","negotiation","expert demonstrations","law firms","rl environments"],"shortTitle":"Legal negotiation trajectories"},{"ref":"MK-31631","title":"Commercial drone footage, imagery and flight logs with job records","headline":"Commercial drone footage, imagery and flight logs with matching job records","description":"Archive of a small commercial drone services business: drone video footage, still imagery, flight logs and the matching job records. The owner says much of it can be delivered already labelled and that it needs no de-identification. Volume, years and formats not yet stated; the owner asks for a NDA before sharing details.","category":"Video & imagery","sector":"Professional services","subsector":"Commercial drone services","dataTypes":["Video","Image","Sensor / telemetry","Tabular / structured"],"useCases":["Vision & perception","Robotics / embodied AI","World models"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":["Drone video footage","Aerial still imagery","Flight logs","Job records"],"formats":[],"delivery":"","labelled":"Partly","labelTypes":["Labels often available (owner-stated)"],"personalData":"None","deidentification":"Not needed","rights":"Owner-held","sample":"On request / under NDA","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Commercial drone services","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["drone","aerial","video","imagery","flight logs","geospatial"],"shortTitle":"Drone footage and flight logs"},{"ref":"MK-31680","title":"Indian seminary: 6 years of bookkeeping, invoices and admin (English and Malayalam)","headline":"Indian seminary: 6 years of bookkeeping, GST invoices and admin records (English, Malayalam)","description":"Six years of financial and administrative records from a small theological seminary in India: accounting-software ledgers, bills and GST invoices, procurement and vendor files and administrative correspondence, in English and Malayalam. Part paper (to be scanned) and part digital; low thousands of documents.","category":"Financial & transactional","sector":"Education","subsector":"Theological seminary","dataTypes":["Documents","Tabular / structured","Email"],"useCases":["Enterprise workflow agents","Domain fine-tuning"],"origin":"First-party company records","regions":["South Asia"],"languages":["English","Malayalam"],"sizeBand":"","count":"Low thousands of documents","volume":"Low thousands of documents","years":"2020–2026","yearsOfHistory":"6","growing":"Yes","systems":["Accounting","Email"],"contents":["Accounting ledgers","GST invoices","Procurement files","Vendor records","Admin correspondence"],"formats":["PDF"],"delivery":"Download","labelled":"Yes","labelTypes":[],"personalData":"Moderate","deidentification":"Planned / on request","rights":"Owner-held","sample":"On request / under NDA","readiness":"Within 2 weeks","stage":"Summary available","scale":"","employees":"","industry":"Theological seminary","tools":[["email","Email"],["accounting","Accounting"]],"span":"6 years","volumeHead":"Low thousands of documents","sizeTB":null,"fromYear":2020,"toYear":2026,"yearsNum":6.0,"tags":["bookkeeping","invoices","india","malayalam","tally","gst","non-profit"],"shortTitle":"Seminary bookkeeping records"},{"ref":"MK-32452","title":"Marketing agency: 5-9 years of email, chat, CRM, calls and projects","headline":"Mid-sized marketing agency offering email, team chat, CRM, accounting, video-call, project, file and HR data over 5-9 years","description":"Operating data of a mid-sized marketing agency covering 5-9 years: Gmail, Slack, HubSpot CRM, QuickBooks accounting, Zoom, file storage, project records and HR. Volumes not yet stated.","category":"Company operations archive","sector":"Marketing, media & publishing","subsector":"Marketing agency","dataTypes":["Email","Chat & messaging","Tabular / structured","Documents"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"5-9","growing":"","systems":["Email (Gmail)","Team chat (Slack)","CRM (HubSpot)","Accounting (QuickBooks)","Video calls (Zoom)","File storage","Project management","HR"],"contents":["Email","Team chat","CRM records","Accounting records","Zoom meetings","Project files","HR records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Mid-size company","employees":"51–200","industry":"Marketing agency","tools":[["slack","Slack"],["hubspot","HubSpot"],["gmail","Gmail"],["quickbooks","QuickBooks"],["zoom","Zoom"],["tickets","Tickets"],["hr","HR & payroll"],["files","File storage"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["marketing agency","hubspot","slack","email","quickbooks","zoom","projects"],"shortTitle":"Mid-sized marketing agency archive"},{"ref":"MK-33078","title":"Remote applied-cryptography startup: 4-year company archive","headline":"Remote-first applied-cryptography startup: 4 years of email, Discord, issues, code and finance records (~1 TB)","description":"Four years of company data from a remote-first, venture-funded applied-cryptography startup working on verifiable, privacy-preserving machine learning. About 1 TB across email (12 inboxes), Discord as the main internal channel, Slack, Notion, Linear, GitHub code and issues, Google Workspace files, the accounting general ledger, e-signed contracts, payroll records and a CRM. Payroll, contracts and customer correspondence are included.","category":"Company operations archive","sector":"Software & internet","subsector":"Applied cryptography / privacy-preserving ML","dataTypes":["Email","Chat & messaging","Code","Documents","Tabular / structured"],"useCases":["Coding models","Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"100 GB-1 TB","count":"~1 TB","volume":"~1 TB; 12 inboxes; 4 years of email and Discord","years":"4 years","yearsOfHistory":"4","growing":"","systems":["Email","Team chat (Discord, Slack)","Notion","Issue tracker (Linear)","GitHub","File storage (Google Workspace)","Accounting (QuickBooks)","E-signature","Payroll","CRM"],"contents":["Email (12 inboxes)","Internal Discord and Slack chat","Notion docs","Linear issues","GitHub code and issues","General ledger","Contracts","Payroll records","CRM"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Export 1-2 weeks","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Applied cryptography / privacy-preserving ML","tools":[["slack","Slack"],["discord","Discord"],["github","GitHub"],["notion","Notion"],["quickbooks","QuickBooks"],["tickets","Tickets"],["email","Email"],["crm","CRM"],["hr","HR & payroll"],["sign","E-signature"],["files","File storage"]],"span":"4 years","volumeHead":"1 TB","sizeTB":1.0,"fromYear":null,"toYear":null,"yearsNum":4.0,"tags":["startup","cryptography","machine learning","code","discord","issue tracker","remote team"],"shortTitle":"Applied-crypto startup archive"},{"ref":"MK-33213","title":"Marketing analytics consultancy: 5 years of audit decks and email","headline":"Boutique marketing analytics consultancy: 5 years of analytics audit decks and email for large ad-campaign clients (~1 TB)","description":"About 1 TB covering 5 years from a small marketing analytics consultancy serving a few large advertising-campaign clients. Contents are analytics audit decks and email. Client deliverables may be covered by engagement terms, which are to be checked.","category":"Company operations archive","sector":"Marketing, media & publishing","subsector":"Marketing analytics consulting","dataTypes":["Documents","Email"],"useCases":["Domain fine-tuning","Enterprise workflow agents","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"100 GB-1 TB","count":"~1 TB","volume":"~1 TB over 5 years","years":"5 years","yearsOfHistory":"5","growing":"","systems":["Email"],"contents":["Analytics audit decks","Client email"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Marketing analytics consulting","tools":[["email","Email"]],"span":"5 years","volumeHead":"1 TB","sizeTB":1.0,"fromYear":null,"toYear":null,"yearsNum":5.0,"tags":["marketing analytics","advertising","consulting","audit decks","email"],"shortTitle":"Ad-campaign analytics audit decks"},{"ref":"MK-34020","title":"Psychiatric practice: ~6 years of staff chat and internal documents","headline":"Psychiatric practice: ~2M internal chat messages plus ~40 GB of internal documents and SOPs over ~6 years","description":"Internal operations data of a psychiatric practice in operation for about 6 years. Offered: ~40 GB of workspace content (internal documents and SOPs, excluding chat) and ~2M internal team-chat messages. Clinical EHR patient records are not offered. Staff chats are likely to mention patients, so chat data needs scrubbing of health information before release.","category":"Company operations archive","sector":"Healthcare & life sciences","subsector":"Psychiatric practice","dataTypes":["Chat & messaging","Documents"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"10-100 GB","count":"~2M chat messages","volume":"~40 GB workspace content (excluding chat); ~2M chat messages","years":"","yearsOfHistory":"6","growing":"","systems":["File storage (Google Workspace)","Team chat (Google Chat)"],"contents":["Internal team chat (~2M messages)","Internal documents","SOPs"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"Required (staff chat mentions patients)","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Psychiatric practice","tools":[["chat","Team chat"],["files","File storage"]],"span":"6 years","volumeHead":"2M chat messages","sizeTB":0.04,"fromYear":null,"toYear":null,"yearsNum":6.0,"tags":["healthcare","psychiatry","clinical operations","team chat","sops","behavioural health"],"shortTitle":"Psychiatric practice staff chat"},{"ref":"MK-34382","title":"Small consulting firm: email, code, meeting recordings, documents and contracts","headline":"2-4 years of data from a small consulting firm: Gmail, GitHub, Google Meet recordings, Drive documents, databases and contracts","description":"Operating data from a small consulting firm covering 2-4 years. The seller lists Gmail, GitHub, Google Meet recordings, Google Drive documents, databases and contracts. Volumes have not been provided yet.","category":"Company operations archive","sector":"Professional services","subsector":"Consulting","dataTypes":["Email","Code","Video","Audio / speech","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Coding models","Speech & audio models","Domain fine-tuning"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"2-4","growing":"","systems":["Gmail","GitHub","Meeting recordings (Google Meet)","Google Drive","Database","Contracts"],"contents":["Email","Source code","Meeting recordings","Documents","Databases","Contracts"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Consulting","tools":[["github","GitHub"],["gmail","Gmail"],["googledrive","Google Drive"],["database","Database"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["consulting","professional services","meeting recordings","code","contracts"],"shortTitle":"Small consultancy work archive"},{"ref":"MK-36039","title":"Cross-border commercial contracts and M&A advisory files (law firm, 13 years)","headline":"13 years of a legal advisory firm's contracts, M&A files and email (~50 TB)","description":"Working files of a legal advisory firm of about 150 people that advises clients worldwide on customer and vendor contracts and M&A. File storage holds about 13 years of email (HTML) and documents (Excel, PDF and PowerPoint work product); the CRM holds about 6 years of invoices, purchase orders and engagement letters. Total about 50 TB across roughly 150 mailboxes. Most material is client work product, so client consent and privilege must be confirmed; export feasibility is not yet confirmed.","category":"Text & documents","sector":"Legal","subsector":"Commercial contracts and M&A advisory","dataTypes":["Documents","Email","Tabular / structured"],"useCases":["Domain fine-tuning","Pre-training","Enterprise workflow agents","Retrieval / RAG"],"origin":"First-party company records","regions":["South Asia"],"languages":[],"sizeBand":"10-100 TB","count":"~50 TB","volume":"~50 TB; ~150 mailboxes","years":"13 years","yearsOfHistory":"13","growing":"","systems":["File storage","CRM","Email"],"contents":["Customer and vendor contracts","M&A advisory files","Email (HTML)","Excel workbooks","PDF documents","PowerPoint decks","Engagement letters","Invoices","Purchase orders"],"formats":["XLSX","PDF","PPTX","HTML email"],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Mid-size company","employees":"51–200","industry":"Commercial contracts and M&A advisory","tools":[["sheet","Excel"],["email","Email"],["crm","CRM"],["files","File storage"]],"span":"13 years","volumeHead":"50 TB","sizeTB":50.0,"fromYear":null,"toYear":null,"yearsNum":13.0,"tags":["legal","contracts","m&a","law firm","email","documents","india","commercial law"],"shortTitle":"Cross-border M&A law files"},{"ref":"MK-36077","title":"High-net-worth mortgage brokerage: ~75K client records and call transcripts","headline":"European high-net-worth mortgage brokerage: ~75K client records with 300-500 data points each, plus call transcripts","description":"Mortgage brokerage serving high-net-worth clients: ~75K individuals with 300-500 data points each, plus advisor call transcripts.","category":"Financial & transactional","sector":"Financial services & insurance","subsector":"Mortgage brokerage","dataTypes":["Tabular / structured","Text"],"useCases":["Domain fine-tuning"],"origin":"First-party company records","regions":["Europe"],"languages":[],"sizeBand":"","count":"~75K clients","volume":"~75K client records; call transcripts","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Mortgage brokerage","tools":[],"span":"","volumeHead":"75K clients","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["mortgage","finance","transcripts","hnw"],"shortTitle":"Mortgage brokerage calls"},{"ref":"MK-36281","title":"US vacation-rental manager: 7-year operating archive","headline":"US vacation-rental management company: 7 years of guest communications, bookings, contracts and email, ~3 TB","description":"Operating data of a small US vacation-rental management company covering 7 years and about 3 TB. Includes thousands of guest communications, email, files, booking records and contracts. Guest records include names, phone numbers, emails and ages.","category":"Company operations archive","sector":"Hospitality & food","subsector":"Vacation-rental management","dataTypes":["Email","Chat & messaging","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"1-10 TB","count":"Thousands of guest communications","volume":"~3 TB; thousands of guest communications","years":"","yearsOfHistory":"7","growing":"","systems":["Email (Gmail)","File storage","Bookings"],"contents":["Guest communications","Email","Booking records","Contracts","Files"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Vacation-rental management","tools":[["gmail","Gmail"],["files","File storage"]],"span":"7 years","volumeHead":"Thousands of guest communications","sizeTB":3.0,"fromYear":null,"toYear":null,"yearsNum":7.0,"tags":["vacation rentals","hospitality","guest communications","bookings","property management"],"shortTitle":"Vacation-rental guest messaging"},{"ref":"MK-36884","title":"Large lighting manufacturer: email, CRM, ERP, CAD and HR records","headline":"Lighting manufacturer offering 5-9 years of email, CRM, ERP, Jira, AutoCAD files, contracts and HR records","description":"Operating data of a lighting manufacturer, covering 5-9 years. Systems named are Outlook email, Zoho CRM, Oracle ERP, Google Drive, Jira and AutoCAD, plus contracts and HR records. Submitted through a referral form; volumes and the link to the company are still to be verified.","category":"Company operations archive","sector":"Manufacturing","subsector":"Lighting manufacturing","dataTypes":["Email","Tabular / structured","Documents","3D / CAD"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"5-9","growing":"","systems":["Email (Outlook)","CRM (Zoho CRM)","ERP (Oracle ERP)","File storage (Google Drive)","Issue tracker (Jira)","CAD (AutoCAD)"],"contents":["Email","CRM records","ERP records","Drive files","Jira issues","AutoCAD drawings","Contracts","HR records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"200+","industry":"Lighting manufacturing","tools":[["jira","Jira"],["email","Outlook"],["googledrive","Google Drive"],["autocad","AutoCAD"],["crm","CRM"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["manufacturing","lighting","erp","crm","autocad","jira","contracts","hr"],"shortTitle":"Lighting maker ERP and CAD"},{"ref":"MK-37211","title":"Custom software studio for crypto and AI startups: 3-year archive","headline":"Custom software studio for crypto and AI startups: 3 years of code, team chat, CRM, accounting and signed contracts","description":"Internal archive of a custom software development studio building for crypto and AI startups, covering about 3 years and ~5 TB. Includes email, team chat (3 years), CRM, accounting, 15+ code projects in Git, cloud documents, 20+ projects on a task board and 40+ e-signed contracts. Much of the code was client work, so ownership of individual repositories and documents is being confirmed.","category":"Company operations archive","sector":"Software & internet","subsector":"Custom software studio (crypto and AI clients)","dataTypes":["Code","Email","Chat & messaging","Documents","Tabular / structured"],"useCases":["Coding models","Enterprise workflow agents","Computer-use & tool-use agents","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"1-10 TB","count":"15+ code projects","volume":"~5 TB; 15+ code projects; 20+ task-board projects; 40+ signed contracts","years":"","yearsOfHistory":"3","growing":"","systems":["Email (Gmail)","Team chat (Slack)","CRM (HubSpot)","Accounting (QuickBooks)","Git (GitHub)","File storage (Google Drive)","Project management (Trello)","E-signature (DocuSign)"],"contents":["Code repositories (15+ projects)","Team chat (3 years)","CRM records","Accounting","Project boards (20+ projects)","Signed contracts (40+)"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Custom software studio","tools":[["slack","Slack"],["github","GitHub"],["trello","Trello"],["hubspot","HubSpot"],["gmail","Gmail"],["googledrive","Google Drive"],["quickbooks","QuickBooks"],["sign","E-signature"]],"span":"3 years","volumeHead":"15 code projects","sizeTB":5.0,"fromYear":null,"toYear":null,"yearsNum":3.0,"tags":["software studio","crypto","web3","ai startups","code","slack","crm","contracts"],"shortTitle":"Crypto AI software studio archive"},{"ref":"MK-37368","title":"IT managed services provider: 20+ year archive","headline":"IT services / MSP with 1K+ customers: 20+ years of email, support tickets, CRM, accounting and issue tracking","description":"Internal systems of an IT services and managed services provider with roughly 1K-9.9K customers and 20+ years in operation: email, team chat, CRM, accounting, support tickets, an issue tracker and a database. The seller says the data is most likely company-owned. Volumes are not stated.","category":"Company operations archive","sector":"Professional services","subsector":"IT managed services provider","dataTypes":["Email","Chat & messaging","Tabular / structured","Text"],"useCases":["Enterprise workflow agents","Computer-use & tool-use agents","Domain fine-tuning","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"20+","growing":"","systems":["Email (Outlook)","Team chat (Slack)","CRM (Salesforce)","Accounting (QuickBooks)","Support tickets","Issue tracker (Jira)","Database"],"contents":["Support tickets","Email","Team chat","CRM records","Accounting","Issue tracker","Database"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"IT managed services provider","tools":[["slack","Slack"],["jira","Jira"],["salesforce","Salesforce"],["email","Outlook"],["quickbooks","QuickBooks"],["database","Database"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["it services","msp","support tickets","helpdesk","crm","jira"],"shortTitle":"20-year IT MSP ticket archive"},{"ref":"MK-37390","title":"Engineering-education company: 6M emails, 120 TB of course media, CRM, tickets and code (2020-2026)","headline":"Engineering edtech: ~6M emails across ~170 mailboxes, ~120 TB of course media and recordings, 1M support tickets, 900K graded submissions, ~4M lines of code","description":"Engineering-education company (mechanical, electrical, civil and electronics courses and projects) used by 70+ automotive companies to train engineers, 90K+ students trained. Email, document and media storage, recordings and student submissions, a 1,391-table database (sales CRM, student CRM, submissions) and code repositories.","category":"Company operations archive","sector":"","subsector":"","dataTypes":["Email","Documents","Video","Database","Tickets","Code"],"useCases":["Enterprise / business agents","Education and tutoring","Customer-service agents","Coding models"],"origin":"","regions":["South Asia"],"languages":[],"sizeBand":"","count":"","volume":"~120 TB; ~6M emails (~170 mailboxes); ~25K student journeys; 1M support tickets; 900K submissions; 40K forum posts; 95 repos (~4M LOC, 28K commits)","years":"2020-2026","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held; student, staff and recording consent to be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"Large organisation","employees":"200+","industry":"","tools":[],"span":"","volumeHead":"120 TB","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":[],"shortTitle":"Engineering-education archive"},{"ref":"MK-37624","title":"Industrial and engineering archives: chip-design verification, manufacturing root-cause, logistics exceptions","headline":"Five sourced archives: chip-design and verification records (21 years), manufacturing fault and root-cause records (16 years), logistics exceptions (7 years), professional-services model QA (9 years), recruitment-tech integration failures","description":"Archives being sourced from several companies (some operating, some insolvent): semiconductor chip-design and verification records (21 years, US), manufacturing fault and root-cause analysis (16 years), logistics exception records (7 years), professional-services model QA (9 years) and recruitment-technology integration failures (4.5 years).","category":"Company operations archive","sector":"Manufacturing","subsector":"Semiconductors / manufacturing / logistics","dataTypes":["Documents","Tabular / structured"],"useCases":["Domain fine-tuning","Enterprise workflow agents"],"origin":"First-party company records","regions":["North America","Europe"],"languages":[],"sizeBand":"","count":"","volume":"","years":"4.5-21 years per archive","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Semiconductors / manufacturing / logistics","tools":[],"span":"21 years","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["semiconductor","verification","root cause","logistics","manufacturing"],"shortTitle":"Industrial engineering archives"},{"ref":"MK-37659","title":"Healthcare company: 10-19 year operations archive (English and Norwegian)","headline":"Healthcare company: 10-19 years of email, Slack, CRM, code, recorded meetings and Drive files in English and Norwegian","description":"Company data from a healthcare business with 10-19 years of history, offered through the owner's corporate-finance advisor. Systems listed: email, team chat, CRM, code repositories, recorded video meetings and file storage. Content is in English and Norwegian. Health-related data is present; de-identification is not yet stated. Volumes are not yet stated.","category":"Company operations archive","sector":"Healthcare & life sciences","subsector":"Healthcare company","dataTypes":["Email","Chat & messaging","Code","Video","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Coding models","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":["English","Norwegian"],"sizeBand":"","count":"","volume":"","years":"10-19 years","yearsOfHistory":"10-19","growing":"","systems":["Email (Gmail)","Team chat (Slack)","CRM (HubSpot)","GitHub","GitLab","Video calls (Meet)","File storage (Google Drive)"],"contents":["Email","Slack chat","HubSpot CRM","GitHub/GitLab code","Meet recordings","Drive files"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"Via authorised intermediary","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Healthcare company","tools":[["slack","Slack"],["github","GitHub"],["gitlab","GitLab"],["hubspot","HubSpot"],["gmail","Gmail"],["googledrive","Google Drive"],["calls","Calls"]],"span":"19 years","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["healthcare","norwegian","crm","code","meeting recordings"],"shortTitle":"Norwegian-English healthcare ops"},{"ref":"MK-37705","title":"Construction sites building whole cities: labour and machine data (Pakistan, UAE, China, US)","headline":"Construction-site labour and machine data from large city-scale builds in Pakistan, UAE, China and US","description":"Data from large construction sites, including whole new cities under construction, in Pakistan, the UAE, China and the US, covering labour activity and construction-machine data. Offered for construction robotics, with labels included. The data is sourced through the owner's partners; volume, fields and formats have not yet been provided.","category":"Physical-world video","sector":"Construction & real estate","subsector":"Large-scale construction sites","dataTypes":["Video","Sensor / telemetry"],"useCases":["Robotics / embodied AI","Vision & perception","World models"],"origin":"Licensed from third parties","regions":["South Asia","Middle East","East Asia","North America"],"languages":[],"sizeBand":"","count":"","volume":"","years":"2026","yearsOfHistory":"","growing":"","systems":[],"contents":["Construction labour data","Construction-machine data","Sites built as complete cities","Multi-country site coverage"],"formats":[],"delivery":"Download","labelled":"Yes","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"On request / under NDA","readiness":"Within 2 weeks","stage":"Summary available","scale":"","employees":"","industry":"Large-scale construction sites","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":2026,"toYear":2026,"yearsNum":null,"tags":["construction","robotics","heavy machinery","labour","job sites","pakistan","uae","china"],"shortTitle":"City-scale construction data"},{"ref":"MK-38090","title":"Independent game studio: pre-AI game-development code and company data","headline":"~4 TB of human-written, pre-AI game-development code from a PC game studio, optionally with team chat and issue tracking","description":"A PC game studio with a released title offers ~4 TB of game-development code written before AI coding tools, or alternatively its whole company dataset including team chat and issue/project tracking. Headcount, years covered, lines of code and rights are not yet confirmed.","category":"Code & software","sector":"Media & entertainment","subsector":"Video game development studio","dataTypes":["Code","Chat & messaging"],"useCases":["Coding models","Enterprise workflow agents"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"1-10 TB","count":"~4 TB code","volume":"~4 TB of game-development code; company data optional","years":"","yearsOfHistory":"","growing":"","systems":["Team chat (Slack)","Issue tracker (Jira/Asana)"],"contents":["Pre-AI game-development code","Team chat","Issue and project tracking"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Video game development studio","tools":[["slack","Slack"],["jira","Jira"],["asana","Asana"]],"span":"","volumeHead":"4 TB code","sizeTB":4.0,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["game development","pre-ai code","human-written code","slack","jira","gamedev"],"shortTitle":"Pre-AI game studio code"},{"ref":"MK-38176","title":"Web-indexing and agent browsing behaviour data","headline":"Web-indexing data and records of how AI agents browse and act on the web, plus environments","description":"Data on how AI agents use the internet and browsers and which actions they take, together with web-indexing data and agent environments. The seller also mentions medical-technology data and professional-network profile data; the source and collection method for any scraped or third-party content has been requested. No volume, time span or breakdown has been provided yet.","category":"Agent traces","sector":"AI & data","subsector":"AI agent web infrastructure","dataTypes":["Agent trajectories","Text","Tabular / structured","RL environment / simulator"],"useCases":["Computer-use & tool-use agents","RL environments","Pre-training","Retrieval / RAG"],"origin":"Mixed","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"","systems":["Web crawler / index","Agent browser logs"],"contents":["Agent browsing behaviour records","Actions taken by agents on websites","Web-indexing data","Agent environments","Medical-technology data (mentioned)","Professional-network profile data (mentioned","source to be specified)"],"formats":["Data + environments"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"AI agent web infrastructure","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["agent traces","web browsing","browser agents","web index","environments","computer use"],"shortTitle":"Agent browsing behaviour"},{"ref":"MK-38205","title":"US franchised convenience store: 10-year store records","headline":"Single US convenience store: 10 years of POS transactions, bookkeeping, inventory audits, payroll and email","description":"Records from one franchised convenience store in the US, covering about 10 years. Includes franchisor POS data (transactions, receipts, daily cash reports), franchise bookkeeping (sales, cost of sales, invoices, shortages), outside inventory-count and audit reports, payroll and the store's email. About 0.1 TB. The franchise agreement may limit sharing of POS and franchisor-system data, so rights need confirming. Payroll contains employee personal data.","category":"Financial & transactional","sector":"Retail & e-commerce","subsector":"Franchised convenience store","dataTypes":["Tabular / structured","Documents","Email"],"useCases":["Forecasting & time series","Enterprise workflow agents","Domain fine-tuning"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"100 GB-1 TB","count":"~10 years of POS records","volume":"~0.1 TB; one store; ~10 years","years":"","yearsOfHistory":"10","growing":"","systems":["POS","Accounting","Payroll (ADP)","Email (Gmail)"],"contents":["POS transactions and receipts","Daily cash reports","Sales, cost of sales, invoices and shortages","Inventory-count and audit reports","Payroll","Email"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Export 1-2 weeks","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Franchised convenience store","tools":[["gmail","Gmail"],["accounting","Accounting"],["shop","Commerce"],["hr","HR & payroll"]],"span":"10 years","volumeHead":"10 years of POS records","sizeTB":0.1,"fromYear":null,"toYear":null,"yearsNum":10.0,"tags":["convenience store","retail","pos","franchise","bookkeeping","inventory audits","payroll"],"shortTitle":"Franchised convenience-store books"},{"ref":"MK-38270","title":"Fractional accounting and FP&A firm: SOPs, close process and financial-model templates","headline":"Fractional accounting/FP&A firm: ~50 SOPs, close schedules, model templates and reporting pipeline","description":"Process IP of a fractional finance firm serving 200+ startup consumer-goods and SaaS clients: ~50 core SOPs for bookkeeping, month-end close and FP&A refined across live engagements, close schedules, three three-statement financial-model templates (consumer goods with subscription and one-time revenue across DTC, marketplace and wholesale; SaaS), and an AI-assisted ETL and reporting pipeline for cohort and retention reporting, with playbooks. The firm's own books may be available anonymised. Client records are excluded.","category":"Expert workflows","sector":"Professional services","subsector":"Fractional accounting and FP&A","dataTypes":["Documents","Tabular / structured","Code"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Evaluation / benchmarks"],"origin":"Expert-created","regions":["North America"],"languages":["English"],"sizeBand":"","count":"~50 SOPs; 3 model templates","volume":"~50 SOPs; 3 model templates; pipeline export","years":"2026","yearsOfHistory":"3.5","growing":"Yes","systems":["Accounting","Project management"],"contents":["Bookkeeping SOPs","Month-end close schedules","Financial-model templates","ETL and reporting pipeline","FP&A playbooks"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"None","deidentification":"Not needed","rights":"Owner-held","sample":"","readiness":"2-6 weeks","stage":"Summary available","scale":"Mid-size company","employees":"51–200","industry":"Fractional accounting and FP&A","tools":[["tickets","Tickets"],["accounting","Accounting"]],"span":"3.5 years","volumeHead":"50 SOPs","sizeTB":null,"fromYear":null,"toYear":2026,"yearsNum":3.5,"tags":["accounting","fp&a","month-end close","sops","financial models","cfo"],"shortTitle":"Fractional finance firm SOPs"},{"ref":"MK-41064","title":"Small newsletter media company: 2-4 year company archive","headline":"Small newsletter media company with a large subscriber base: email, chat, CRM, recordings, docs, database and contracts","description":"Company data from a small media and newsletter business with a subscriber base in the 100K-999K range and 2-4 years of history. Systems listed: email, team chat, CRM, video-call recordings, file storage, Notion, a database and e-signed contracts. Subscriber-level engagement data is likely personal. Volumes are not yet stated.","category":"Company operations archive","sector":"Marketing, media & publishing","subsector":"Newsletter publisher","dataTypes":["Email","Chat & messaging","Documents","Video","Tabular / structured"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Recommendation & personalisation","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"2-4 years","yearsOfHistory":"2-4","growing":"","systems":["Email (Gmail)","Team chat (Slack)","CRM","Video calls (Meet)","File storage (Google Drive)","Notion","Database","E-signature"],"contents":["Email","Slack chat","CRM","Meet recordings","Drive files","Notion pages","Database incl. subscriber engagement","Contracts"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Newsletter publisher","tools":[["slack","Slack"],["gmail","Gmail"],["notion","Notion"],["googledrive","Google Drive"],["crm","CRM"],["calls","Calls"],["database","Database"],["sign","E-signature"]],"span":"4 years","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["media","newsletter","subscribers","engagement","crm","notion"],"shortTitle":"Newsletter media company archive"},{"ref":"MK-41267","title":"200K human-made illustrations, artworks, vector, CAD and schematics","headline":"~200K human-made illustrations, artworks, photos, vector graphics, video, CAD, schematics and infographics from contracted artists; partly labelled, growing","description":"Original human-made visual works from a network of contracted artists in India: illustrations, artwork, photos, vector/SVG, video, CAD, schematics and infographics. Partly labelled; more added monthly.","category":"Video & imagery","sector":"Marketing, media & publishing","subsector":"Illustration and artwork","dataTypes":["Image","Video","3D / CAD"],"useCases":["Pre-training","Vision & perception"],"origin":"Expert-created","regions":["South Asia"],"languages":[],"sizeBand":"","count":"~200K works","volume":"~200K works","years":"2025–2026","yearsOfHistory":"1","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"Partly","labelTypes":[],"personalData":"None","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Illustration and artwork","tools":[],"span":"1 year","volumeHead":"200K works","sizeTB":null,"fromYear":2025,"toYear":2026,"yearsNum":1.0,"tags":["illustration","artwork","vector","images","creative"],"shortTitle":"Human-made illustrations and CAD"},{"ref":"MK-42389","title":"B2B outbound agency: 4-5 years of campaign histories, client negotiations and AI chat logs (contact databases excluded)","headline":"UAE-based outbound sales agency: Pipedrive CRM, client negotiation email, hundreds of thousands of campaign sequences (finance, PE, oil and gas, staffing) and 3 years of internal AI chat logs","description":"Operating records of a B2B cold-email and LinkedIn outreach agency: CRM, client negotiation threads, campaign histories with reply outcomes across finance, private equity, oil and gas and staffing clients, records from a partial company sale, and three years of internal AI-assistant chat logs. Third-party contact databases are excluded.","category":"Company operations archive","sector":"Professional services","subsector":"Sales agency","dataTypes":["Email","Tabular / structured","Chat & messaging"],"useCases":["Sales agents","Enterprise workflow agents"],"origin":"First-party company records","regions":["Middle East"],"languages":[],"sizeBand":"","count":"","volume":"Hundreds of thousands of campaigns; 2,000 sending inboxes; total stated ~30 TB (to verify)","years":"4-5 years","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Medium","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Sales agency","tools":[["crm","Pipedrive"]],"span":"5 years","volumeHead":"Hundreds of thousands of campaigns","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["outbound","sales","crm","campaigns","ai chat logs"],"shortTitle":"B2B outbound agency archive"},{"ref":"MK-42516","title":"Esports competition platform history and neobank transactions","headline":"~8 years of esports competition platform data (matches, results, wagering, payouts) plus a neobank transaction dataset","description":"Two assets from operators that are no longer generating data. (1) About 8 years of history from an esports competition platform that has shut down: matches, results, player performance, and wagering and payout history, which the seller still owns. (2) Separately, a consumer neobank transaction dataset; the neobank's brand and products were sold but the data was not, and what can be shared is subject to a review of its terms. The seller is open to exclusive or non-exclusive sale and has offered a sample schema under NDA while the inventory is prepared.","category":"Financial & transactional","sector":"Consumer services","subsector":"Esports competition platform; consumer neobank","dataTypes":["Tabular / structured"],"useCases":["Forecasting & time series","Domain fine-tuning","Pre-training"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"~8 years of platform history","volume":"","years":"~8 years","yearsOfHistory":"8","growing":"No","systems":["Database"],"contents":["Esports matches","Match results","Player performance","Wagering history","Payout history","Consumer neobank transactions"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"On request / under NDA","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Esports competition platform","tools":[["database","Database"]],"span":"8 years","volumeHead":"8 years of platform history","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":8.0,"tags":["esports","gaming","wagering","payouts","player performance","neobank","transactions","fintech"],"shortTitle":"Esports and neobank data"},{"ref":"MK-42642","title":"Heavy mining-equipment multimodal sensor data","headline":"Heavy mining-equipment sensor data (RGB, lidar, CAN bus, voice) from US mines, monthly supply","description":"Multimodal sensor data from seller-installed recorders on excavators, wheel loaders, haul trucks and bulldozers at 7 US mine sites: RGB video, lidar, anonymised location, decoded J1939 CAN bus signals, operator voice and the machine's safety state, all on one shared clock with one continuous video per session and events timestamped into it. Today only a short pilot from 3 machines exists; full collection starts Q4 2026 with ~100 units deployed before January and a projected ~60,000 hours per month from ~110 collectors, delivered in monthly batches. Labels are generated automatically and audited manually; metadata flags low-quality data, packet loss is about 0.01%, and voice sometimes needs scrubbing. GPS has been replaced with locally bounded coordinates; some streams are owned and others licensed to the seller in perpetuity for resale and training.","category":"Industrial & sensor","sector":"Energy, mining & utilities","subsector":"Mining heavy equipment","dataTypes":["Video","Sensor / telemetry","Audio / speech","Geospatial"],"useCases":["Robotics / embodied AI","World models","Vision & perception","Forecasting & time series"],"origin":"Purpose-collected (consented capture)","regions":["North America"],"languages":[],"sizeBand":"","count":"~60K hours/month (projected)","volume":"Projected ~60,000 hours/month from ~110 collectors","years":"2026","yearsOfHistory":"","growing":"Yes","systems":["Onboard recorders","CAN bus"],"contents":["RGB video","Lidar","Anonymised location (local coordinates)","Decoded CAN bus (J1939)","Operator voice","Machine safety state","Quality-flag metadata","Auto labels with manual audit"],"formats":["MP4","MP3","JSON","MCAP (on request)"],"delivery":"Monthly batches; Cloud bucket (GCP/AWS, expected)","labelled":"Yes","labelTypes":["Automatic labels","Manual label audit","Outcomes","Metadata"],"personalData":"Moderate","deidentification":"Done by owner","rights":"Owner-held, conditions to confirm","sample":"Not available","readiness":"Collected to order","stage":"Collected to order","scale":"","employees":"","industry":"Mining heavy equipment","tools":[],"span":"","volumeHead":"60K hours/month","sizeTB":null,"fromYear":2026,"toYear":null,"yearsNum":null,"tags":["mining","heavy equipment","lidar","can bus","excavator","haul truck","sensor fusion","autonomy"],"shortTitle":"Mining-equipment sensor data"},{"ref":"MK-42960","title":"Professional users' AI requests and prompt rewrites","headline":"10M+ AI requests from non-technical professionals, with how their first prompts fall short","description":"Over 10 million AI requests submitted by non-technical professionals (legal, sales, marketing, founders, accountants) through the seller's own prompt-optimisation product, which has close to 400K monthly users. Each record captures what the professional wanted from AI and how their initial prompt fell short of it. Date range, delivery format and confirmation that user terms allow licensing have not yet been provided.","category":"Human preference / evaluation","sector":"Software & internet","subsector":"AI prompt-optimisation tool","dataTypes":["Text"],"useCases":["SFT / instruction tuning","Preference / RLHF","Evaluation / benchmarks"],"origin":"Platform / user-generated","regions":["Not stated"],"languages":[],"sizeBand":"","count":"10M+ AI requests","volume":"10M+ AI requests; ~400K monthly users","years":"","yearsOfHistory":"","growing":"Yes","systems":["Product database"],"contents":["User-written AI requests / prompts","Initial prompt versus optimised intent","Requests from legal professionals","Requests from sales and marketing users","Requests from founders and accountants"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"AI prompt-optimisation tool","tools":[["database","Database"]],"span":"","volumeHead":"10M AI requests","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["prompts","user intent","prompt optimisation","professionals","instruction tuning","llm usage"],"shortTitle":"Professional AI prompt rewrites"},{"ref":"MK-43509","title":"Diagnostic reasoning tasks across medical specialties","headline":"Diagnostic reasoning tasks and records across multiple medical specialties","description":"Diagnostic reasoning tasks and records spanning multiple healthcare specialties, offered by a healthcare-focused data provider. The seller states it has access to the data rather than describing it as its own records. Volume, format, time span, de-identification and rights basis have not yet been provided.","category":"Healthcare & life sciences","sector":"Healthcare & life sciences","subsector":"Clinical diagnostic reasoning","dataTypes":["Text"],"useCases":["Domain fine-tuning","Evaluation / benchmarks","SFT / instruction tuning"],"origin":"","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":["Diagnostic reasoning tasks","Diagnostic records","Coverage across multiple medical specialties"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Clinical diagnostic reasoning","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["healthcare","diagnosis","medical reasoning","clinical","specialties","evaluation"],"shortTitle":"Medical diagnostic reasoning"},{"ref":"MK-44915","title":"Egocentric manufacturing video, automotive parts plants (to order)","headline":"Egocentric + multi-angle factory video from automotive-parts plants in India (to order, ~50-hour pilot)","description":"Planned capture in the seller's own automotive-parts manufacturing plants (suspension parts, transmission parts, CNC-machined parts, die castings): operator head-camera video synchronised with fixed multi-angle cameras. Tasks include CNC loading and unloading, deburring, gauging, trimming die castings, sub-assembly, kitting, material handling, inspection and packing, with task/step labels and optional hand/object annotations. A ~50-hour pilot on 2-3 stations can start within about two weeks, then scale across both plants; written operator consent is being collected under India's data-protection law and customer filming permissions are being cleared.","category":"Physical-world video","sector":"Manufacturing","subsector":"Automotive parts manufacturing","dataTypes":["Video"],"useCases":["Robotics / embodied AI","World models","Vision & perception"],"origin":"Purpose-collected (consented capture)","regions":["South Asia"],"languages":[],"sizeBand":"","count":"~50-hour pilot (to order)","volume":"~50-hour pilot, then scale-up across two plants","years":"Not yet captured","yearsOfHistory":"","growing":"Yes","systems":["Head cameras","Fixed multi-angle cameras"],"contents":["Operator head-camera video","Synchronised fixed multi-angle video","CNC loading/unloading","Deburring and gauging","Die-casting trimming","Sub-assembly and kitting","Material handling","Inspection and packing","Task/step labels"],"formats":["Egocentric + multi-angle video"],"delivery":"","labelled":"Yes","labelTypes":["Task / step labels","Optional hand / object annotations"],"personalData":"Moderate","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Collected to order","stage":"Collected to order","scale":"","employees":"","industry":"Automotive parts manufacturing","tools":[],"span":"","volumeHead":"~50-hour pilot","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["egocentric","manufacturing","factory","cnc","die casting","manipulation","india","automotive"],"shortTitle":"Auto-parts plant egocentric video"},{"ref":"MK-45156","title":"Coding bootcamp: 6,000+ recorded sales calls and 86K processed leads","headline":"Coding bootcamp: 6,000+ recorded sales calls, 86K processed leads and video lessons","description":"Records of a coding bootcamp running for 4+ years: a CRM with 86,000 processed leads (~4,000 a month) including emails, phone and video calls, 6,000+ recorded closing sales calls, in-house video lessons, marketing campaigns and team and student chat. Recording consent and CRM terms for prospects still need confirming.","category":"Speech & conversation","sector":"Education","subsector":"Coding bootcamp","dataTypes":["Audio / speech","Video","Email","Tabular / structured","Chat & messaging"],"useCases":["Speech & audio models","Enterprise workflow agents","Domain fine-tuning"],"origin":"First-party company records","regions":["North America"],"languages":["English"],"sizeBand":"1-10 TB","count":"6,000+ sales calls","volume":"~1 TB; 86K leads; 6,000+ sales-call recordings","years":"2022–2026","yearsOfHistory":"4","growing":"Yes","systems":["CRM","File storage","Team chat"],"contents":["Recorded sales calls","CRM lead records","Video lessons","Marketing campaigns","Team and student chat"],"formats":[],"delivery":"","labelled":"Partly","labelTypes":["Deal outcomes"],"personalData":"High","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Within 2 weeks","stage":"Summary available","scale":"Mid-size company","employees":"51–200","industry":"Coding bootcamp","tools":[["chat","Team chat"],["crm","CRM"],["files","File storage"]],"span":"4 years","volumeHead":"6K sales calls","sizeTB":1.0,"fromYear":2022,"toYear":2026,"yearsNum":4.0,"tags":["sales calls","crm","bootcamp","education","leads","video lessons"],"shortTitle":"Bootcamp sales calls"},{"ref":"MK-45556","title":"European non-profit choir: 10 years of emails, reports, scores and concert recordings","headline":"Ten years of records from a European non-profit choir: emails on public funding and events, financial reports, assembly notes, scores and concert audio and video","description":"About ten years of records from a non-profit choir in Europe. Includes emails on government funding, events and logistics (stated as ~3K to ~10K emails before parsing), annual and financial reports, assembly notes, musical scores and concert audio and video. Total size is stated as 15-20 GB of files. The emails include correspondence with individuals.","category":"Text & documents","sector":"Media & entertainment","subsector":"Non-profit choir","dataTypes":["Email","Documents","Audio / speech","Video","Music scores"],"useCases":["Domain fine-tuning","Retrieval / RAG","Music & audio models"],"origin":"First-party company records","regions":["Europe"],"languages":[],"sizeBand":"10-100 GB","count":"~3K-10K emails","volume":"15-20 GB of files; ~3K-10K emails (estimates vary); reports, scores and concert recordings","years":"","yearsOfHistory":"10","growing":"","systems":["Email","File storage"],"contents":["Emails on government funding, events and logistics","Annual and financial reports","Assembly notes","Musical scores","Concert audio and video"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Non-profit choir","tools":[["email","Email"],["files","File storage"]],"span":"10 years","volumeHead":"~3K-10K emails","sizeTB":0.02,"fromYear":null,"toYear":null,"yearsNum":10.0,"tags":["non-profit","choir","music","concert recordings","scores","email","europe"],"shortTitle":"Non-profit choir emails and audio"},{"ref":"MK-46150","title":"Garment-factory human demonstrations for robotics (to order)","headline":"Human demonstrations of garment making (handling, cutting, sewing, assembly, finishing, QC, packing) in Bangladeshi factories: ~60 h pilot, scaling to 500+ participants","description":"Recorded human demonstrations across garment production in Bangladeshi factories: fabric handling, cutting, sewing, assembly, finishing, quality control and packing. ~60-hour pilot, scalable to 500+ participants; access to thousands of factories.","category":"Physical-world video","sector":"Manufacturing","subsector":"Garments","dataTypes":["Video"],"useCases":["Robotics / embodied AI"],"origin":"Purpose-collected (consented capture)","regions":["South Asia"],"languages":[],"sizeBand":"","count":"~60 hours (pilot)","volume":"~60 h pilot; scalable","years":"2026","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"Collected to order","stage":"Collected to order","scale":"","employees":"","industry":"Garments","tools":[],"span":"","volumeHead":"60 hours","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["garments","factory","demonstrations","robotics"],"shortTitle":"Garment-factory demonstrations"},{"ref":"MK-46558","title":"Civil and structural engineering practice: ~500 public-works projects over 25 years","headline":"Civil and structural engineering archive: ~500 public-works projects over 25 years","description":"The technical archive of a civil and structural engineering practice that works mainly for public bodies: ~500 projects (roads, ski lifts, buildings and more) with design documentation, structural calculations and work-progress reports, covering 25 years.","category":"Industrial & sensor","sector":"Construction & real estate","subsector":"Civil and structural engineering","dataTypes":["Documents","Tabular / structured","3D / CAD"],"useCases":["Domain fine-tuning","Enterprise workflow agents","Retrieval / RAG"],"origin":"First-party company records","regions":["Europe"],"languages":["Italian"],"sizeBand":"1-10 TB","count":"~500 projects","volume":"~5 TB; ~500 projects","years":"2001–2026","yearsOfHistory":"25","growing":"Yes","systems":["File storage","Spreadsheets"],"contents":["Design documentation","Structural calculations","Work-progress reports","Public-works projects"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Within 2 weeks","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Civil and structural engineering","tools":[["sheet","Spreadsheets"],["files","File storage"]],"span":"25 years","volumeHead":"500 projects","sizeTB":5.0,"fromYear":2001,"toYear":2026,"yearsNum":25.0,"tags":["civil engineering","structural calculations","public works","infrastructure","italy"],"shortTitle":"Civil engineering practice"},{"ref":"MK-46710","title":"Specialty coffee brand: 8 years of e-commerce and warehouse ERP records","headline":"Coffee brand: ~600K e-commerce rows and ~500K ERP rows (warehouse, inward and outward goods, fulfilment) over ~8 years","description":"Coffee brand with about eight years of business records: e-commerce store data (~600K rows) and warehouse ERP data (~500K rows covering inward and outward goods and fulfilment). Customer contact details removed before delivery.","category":"Company operations archive","sector":"Retail & e-commerce","subsector":"Coffee brand","dataTypes":["Tabular / structured"],"useCases":["Enterprise workflow agents","Forecasting & time series"],"origin":"First-party company records","regions":["South Asia"],"languages":[],"sizeBand":"","count":"","volume":"~600K e-commerce rows; ~500K ERP rows","years":"~8 years","yearsOfHistory":"","growing":"","systems":["E-commerce","ERP"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"Planned / on request","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Coffee brand","tools":[["erp","ERP"],["shop","Commerce"]],"span":"8 years","volumeHead":"600K e-commerce rows","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["coffee","e-commerce","erp","fulfilment"],"shortTitle":"Specialty coffee brand records"},{"ref":"MK-46970","title":"Hardware and software startup: software revisions and hardware design files since 2018","headline":"Startup operating since 2018 with years of software revision history and hardware design information","description":"Data from a small startup operating since 2018. The source describes years of software revision history and years of hardware design information. File counts, formats and volumes have not been provided yet.","category":"Code & software","sector":"Other","subsector":"Hardware and software startup","dataTypes":["Code","CAD / design files"],"useCases":["Coding models","Engineering & CAD models"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"2018","yearsOfHistory":"","growing":"","systems":[],"contents":["Software revision history","Hardware design information"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Hardware and software startup","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":2018,"toYear":null,"yearsNum":null,"tags":["startup","hardware design","software revisions","code","engineering"],"shortTitle":"Hardware startup design history"},{"ref":"MK-47811","title":"AI coding-tool startup: 6 years of CRM, demo recordings, Jira, Slack, Notion roadmap and product analytics (~10 TB)","headline":"Legacy-code AI copilot company: CRM with demo recordings, Drive, email, Slack, Jira, a Notion roadmap from pre-LLM onward, finance and product-usage analytics","description":"Operating records of an AI coding-copilot company focused on legacy codebases: CRM with recorded product demos, shared drive (fundraising, recruiting, legal), email, team chat, Jira, a product roadmap running from before LLMs to today, accounting and product-usage analytics from enterprise customers. Payroll, employee records and any scraped contact lists are excluded.","category":"Company operations archive","sector":"Software & SaaS","subsector":"Developer tools","dataTypes":["Email","Team chat","Tickets","Documents","Call recordings","Tabular / structured"],"useCases":["Enterprise workflow agents","Coding models"],"origin":"First-party company records","regions":["North America","South Asia"],"languages":[],"sizeBand":"1-10 TB","count":"","volume":"~10 TB stated","years":"~6 years","yearsOfHistory":"","growing":"","systems":["HubSpot","Google Drive","Outlook","Slack","Jira","Notion","QuickBooks","Mixpanel/PostHog"],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Medium","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Developer tools","tools":[["slack","Slack"],["jira","Jira"],["hubspot","HubSpot"],["email","Outlook"],["notion","Notion"],["googledrive","Google Drive"],["quickbooks","QuickBooks"],["mixpanel","Mixpanel"]],"span":"6 years","volumeHead":"10 TB stated","sizeTB":10.0,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["startup archive","developer tools","jira","roadmap","demos"],"shortTitle":"AI coding-tool startup archive"},{"ref":"MK-48123","title":"Small legal-sector company: 2-4 years of documents and recordings","headline":"Small legal-sector company: 2-4 years of email, accounting, video-call recordings, Drive files and e-signed contracts","description":"Company data from a small legal-sector business serving hundreds of customers, with 2-4 years of history. Systems listed: email, accounting, video-call recordings, file storage and contracts with e-signatures. Includes client legal documents and recordings. Working languages are Hindi and English. Ownership is described as most likely held by the company; volumes are not yet stated.","category":"Company operations archive","sector":"Legal","subsector":"Legal services","dataTypes":["Email","Documents","Audio / speech","Video","Tabular / structured"],"useCases":["Domain fine-tuning","Speech & audio models","Enterprise workflow agents","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":["English","Hindi"],"sizeBand":"","count":"","volume":"","years":"2-4 years","yearsOfHistory":"2-4","growing":"","systems":["Email (Gmail)","Accounting","Video calls (Zoom)","File storage (Google Drive)","E-signature"],"contents":["Email","Accounting records","Zoom recordings","Client legal documents","Contracts and e-signatures"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Timing to confirm","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Legal services","tools":[["gmail","Gmail"],["googledrive","Google Drive"],["zoom","Zoom"],["accounting","Accounting"],["sign","E-signature"]],"span":"4 years","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["legal","legal services","contracts","call recordings","hindi","english"],"shortTitle":"Legal-sector docs and recordings"},{"ref":"MK-48367","title":"UAE logistics company: complete internal records","headline":"UAE-based global logistics company: vendors, customers, invoices, revenue, employees and all email","description":"Complete internal records of a family-owned global logistics company based in the United Arab Emirates, about 5 years old: vendor list, customer list, invoices and invoice details, revenue records, employee list and all company email, possibly also business messaging-app conversations. Around 3,000 invoices/customers are stated. Offered through a family member of the owner.","category":"Company operations archive","sector":"Logistics, distribution & wholesale","subsector":"Global logistics / freight services","dataTypes":["Email","Tabular / structured","Documents","Chat & messaging"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Pre-training","Retrieval / RAG"],"origin":"First-party company records","regions":["Middle East"],"languages":[],"sizeBand":"","count":"~3,000 invoices/customers","volume":"~3,000 invoices/customers","years":"2026","yearsOfHistory":"5","growing":"Yes","systems":["Email","Accounting","Business messaging"],"contents":["Vendor list","Customer list","Invoices and invoice details","Revenue records","Employee list","All company email","Business messaging conversations (possible)"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Global logistics / freight services","tools":[["chat","Team chat"],["email","Email"],["accounting","Accounting"]],"span":"5 years","volumeHead":"3K invoices/customers","sizeTB":null,"fromYear":null,"toYear":2026,"yearsNum":5.0,"tags":["logistics","freight","invoices","email","vendors","customers","uae","middle east"],"shortTitle":"UAE logistics company records"},{"ref":"MK-48429","title":"Production cross-platform consumer app: coding and app-use RL environment","headline":"Sealed copy of a live consumer subscription app with real production bugs, migration tasks and app-use goals, hidden-test rewards","description":"RL environment built from a sealed copy of a live consumer subscription app (~700K installs, live since 2025) from a European developer: React Native/Expo client, real-time TypeScript backend, admin dashboard, plus the native SwiftUI version it migrated from. Tasks include real production bugs from 16 months of history (~2K commits) with the original fixes as references, SwiftUI to React Native migration tasks, iOS/Android app-use goals and two-agent pairing tasks. Rewards come from hidden tests and backend state; delivered as SWE-bench-style Docker images with a Gymnasium wrapper. Offered as a pilot, scalable to several hundred tasks. No real user data (backend data is synthetic). Third-party licences are being checked. Non-exclusive, time-limited licence preferred.","category":"RL environment / benchmark","sector":"Software & internet","subsector":"Consumer subscription mobile app","dataTypes":["Code"],"useCases":["Coding models","Computer-use & tool-use agents"],"origin":"First-party company records","regions":["Europe"],"languages":[],"sizeBand":"","count":"~2K commits","volume":"Pilot task set, scalable to several hundred tasks; 16 months of production history (~2K commits)","years":"2025–2026","yearsOfHistory":"1","growing":"","systems":["Git","Database (Convex)"],"contents":["React Native/Expo client","TypeScript backend","Admin dashboard","Native SwiftUI predecessor app","Production bug-fix tasks with reference fixes","SwiftUI to React Native migration tasks","iOS/Android app-use goals","Two-agent pairing tasks"],"formats":["Docker images","Gymnasium wrapper"],"delivery":"","labelled":"Yes","labelTypes":["Hidden tests","Backend-state rewards","Reference fixes"],"personalData":"None","deidentification":"No user data (synthetic backend data)","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Export about a month","stage":"Summary available","scale":"","employees":"","industry":"Consumer subscription mobile app","tools":[["docker","Docker"],["database","Database"]],"span":"1 year","volumeHead":"2K commits","sizeTB":null,"fromYear":2025,"toYear":2026,"yearsNum":1.0,"tags":["rl environment","swe","mobile app","react native","swiftui","migration","app-use agent","hidden tests"],"shortTitle":"Live consumer app SWE/app-use env"},{"ref":"MK-48719","title":"Cannabis retail and ecommerce business: orders, POS, CRM and support","headline":"Cannabis retailer with tens of thousands of customers: 5-9 years of email, chat, CRM, accounting, support, wiki and POS orders","description":"Operating data of a cannabis retail and ecommerce business with tens of thousands of customers, covering 5-9 years: email, chat, CRM, accounting, customer support, file storage, wiki and order/POS records. Customer purchase records for a regulated product are included. Volumes not yet stated.","category":"Company operations archive","sector":"Retail & e-commerce","subsector":"Cannabis retail and ecommerce","dataTypes":["Email","Chat & messaging","Tabular / structured","Documents"],"useCases":["Enterprise workflow agents","Forecasting & time series","Domain fine-tuning"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"Tens of thousands of customers","years":"","yearsOfHistory":"5-9","growing":"","systems":["Email","Team chat","CRM","Accounting","Customer support","File storage","Wiki","POS"],"contents":["Email","Chat","CRM records","Accounting records","Support tickets","Files","Wiki pages","Orders and POS transactions"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Cannabis retail and ecommerce","tools":[["chat","Team chat"],["tickets","Tickets"],["email","Email"],["crm","CRM"],["wiki","Docs & wiki"],["accounting","Accounting"],["shop","Commerce"],["files","File storage"]],"span":"","volumeHead":"Tens of thousands of customers","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["cannabis","retail","ecommerce","pos","crm","customer support","orders"],"shortTitle":"Cannabis retailer orders and POS"},{"ref":"MK-49501","title":"Performance-marketing agency for insurance and finance clients: 5 years of campaigns","headline":"Performance-marketing agency for insurance and finance: 5 years of campaigns and finances","description":"Five years of records from a small performance-marketing agency running paid traffic for insurance and financial companies: campaign results (cost per acquisition, leads generated), ad creatives, client payments and conversations, invoices and company finances.","category":"Company operations archive","sector":"Marketing, media & publishing","subsector":"Performance-marketing agency","dataTypes":["Tabular / structured","Email","Image","Documents"],"useCases":["Domain fine-tuning","Enterprise workflow agents"],"origin":"First-party company records","regions":["North America"],"languages":["English"],"sizeBand":"100 GB-1 TB","count":"~500 GB","volume":"~500 GB","years":"2021–2026","yearsOfHistory":"5","growing":"Yes","systems":["File storage","Email"],"contents":["Campaign CPA and lead results","Ad creatives","Client conversations","Invoices and payments"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"Within 2 weeks","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Performance-marketing agency","tools":[["email","Email"],["files","File storage"]],"span":"5 years","volumeHead":"500 GB","sizeTB":0.5,"fromYear":2021,"toYear":2026,"yearsNum":5.0,"tags":["advertising","performance marketing","insurance","lead generation","agency"],"shortTitle":"Insurance marketing agency"},{"ref":"MK-49765","title":"Agent integrity traces (reward-hacking and honesty)","headline":"Agent integrity traces: sandboxed tasks testing whether agents cheat, reward-hack or misreport","description":"Agent traces from sandboxed tasks in mathematics, coding, creative writing and self-reporting of resource use and task completion, in which agents are told not to cheat but given opportunities to (accessing restricted answers, consulting other models, misreporting results). Records include agent conversations, maths solutions, software changes, creative writing, self-reported performance and tool-use logs, showing whether agents take shortcuts and report honestly. Planned annotations cover task completion, attempted or taken prohibited shortcuts and accuracy of self-reports, backed by independently checked outcomes. Generated from September 2026 in sandboxes and task designs the seller builds; the seller notes coverage starts with controlled scenarios and borderline cases may need human review.","category":"Agent traces","sector":"AI & data","subsector":"Agent safety / reward hacking","dataTypes":["Agent trajectories","Text","Code"],"useCases":["Evaluation / benchmarks","Preference / RLHF","RL environments","Computer-use & tool-use agents"],"origin":"Generated / simulated","regions":["North America"],"languages":["English"],"sizeBand":"","count":"","volume":"Not stated (built to order)","years":"2026","yearsOfHistory":"","growing":"Yes","systems":["Sandbox environments"],"contents":["Agent conversations","Mathematical solutions","Software changes","Creative writing outputs","Self-reported performance and resource use","Tool-use logs","Shortcut / cheating opportunities","Independently checked outcomes"],"formats":["JSON","JSONL"],"delivery":"File share","labelled":"Partly","labelTypes":["Planned: task completion","Planned: attempted / taken prohibited shortcuts","Planned: accuracy of self-reports"],"personalData":"None","deidentification":"Not needed","rights":"Owner-held","sample":"Available","readiness":"Within 2 weeks","stage":"Collected to order","scale":"","employees":"","industry":"Agent safety / reward hacking","tools":[],"span":"","volumeHead":"Not stated","sizeTB":null,"fromYear":2026,"toYear":null,"yearsNum":null,"tags":["agents","reward hacking","honesty","alignment","traces","evals","safety","sandbox"],"shortTitle":"Agent integrity traces"},{"ref":"MK-49846","title":"Software consultancy: projects, tickets and engineering docs","headline":"Small software consultancy: client projects, tickets and engineering documentation","description":"Client project records, tickets and engineering documentation from a very small, founder-run software and IT services business. Volumes, formats and time span have not been provided.","category":"Code & software","sector":"Software & internet","subsector":"Software consultancy / IT services","dataTypes":["Documents","Text","Tabular / structured"],"useCases":["Coding models","Enterprise workflow agents","Domain fine-tuning"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"","systems":["Issue tracker","File storage"],"contents":["Client project records","Tickets","Engineering documentation"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Software consultancy / IT services","tools":[["tickets","Tickets"],["files","File storage"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["software consultancy","it services","tickets","engineering docs","client projects"],"shortTitle":"Software consultancy tickets"},{"ref":"MK-51206","title":"Legal-compliance RL data and evaluations","headline":"RL data and proprietary evaluations for training models to know and follow laws and regulations","description":"RL data plus proprietary evaluations aimed at aligning models to behave lawfully, i.e. to know and adhere to laws and regulations. The seller describes the RL data as highly unique. Volume, jurisdictions, formats and delivery have not yet been provided.","category":"RL environment / benchmark","sector":"AI & data","subsector":"Legal-compliance alignment","dataTypes":["Text"],"useCases":["RL environments","Evaluation / benchmarks","Preference / RLHF"],"origin":"","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":["RL training data for lawful behaviour","Proprietary compliance evaluations","Law and regulation adherence tasks"],"formats":["RL data","Evals"],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Legal-compliance alignment","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["alignment","legal compliance","regulation","reinforcement learning","evals","safety"],"shortTitle":"Legal-compliance RL data"},{"ref":"MK-51426","title":"Events-ticketing startup: codebase, support tickets and transactions","headline":"Ticketing startup: codebase first, plus support tickets, transaction records and documents (~1-1.5 years)","description":"Event-ticketing software company offering its codebase first, with support tickets, transaction records and internal documents available later. Part of the data is partner-owned and excluded.","category":"Code & software","sector":"Software & internet","subsector":"Event ticketing","dataTypes":["Code","Tabular / structured","Documents"],"useCases":["Coding models"],"origin":"First-party company records","regions":["North America","South Asia"],"languages":[],"sizeBand":"","count":"","volume":"","years":"~1-1.5 years","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Event ticketing","tools":[],"span":"1.5 years","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["ticketing","code","support tickets"],"shortTitle":"Event-ticketing startup"},{"ref":"MK-52708","title":"Specialty-clinic operational workflows","headline":"Operational and workflow data from 15+ US specialty clinics, prepared for AI training","description":"Operational and workflow data drawn from more than 15 specialty medical clinics in Florida, described by the seller as built for AI training. The seller states it contains no patient information. Record counts, systems, formats and time span have not yet been provided.","category":"Healthcare & life sciences","sector":"Healthcare & life sciences","subsector":"Specialty clinic operations","dataTypes":["Tabular / structured"],"useCases":["Enterprise workflow agents","Domain fine-tuning"],"origin":"","regions":["North America"],"languages":[],"sizeBand":"","count":"15+ clinics","volume":"15+ clinics","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":["Clinic operational records","Clinic workflow data","Coverage across 15+ specialty clinics"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Specialty clinic operations","tools":[],"span":"","volumeHead":"15 clinics","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["healthcare","clinics","operations","workflow","specialty care","florida"],"shortTitle":"Specialty-clinic workflows"},{"ref":"MK-54208","title":"Consumer wearable health app: VO2max, HRV, sleep and activity data since 2019","headline":"Consumer health app with 1M+ lifetime installs ingesting wearable metrics: VO2max, resting heart rate, HRV, steps and sleep","description":"Data from a consumer health-tech app operating since 2019 with over 1M lifetime installs. The app ingests wearable health data from its users, including VO2max, resting heart rate, heart-rate variability, steps and sleep. This is personal health data; consent terms for licensing and de-identification approach are not yet stated.","category":"Healthcare & life sciences","sector":"Healthcare & life sciences","subsector":"Consumer wearable health app","dataTypes":["Sensor / telemetry","Tabular / structured"],"useCases":["Forecasting & time series","Domain fine-tuning"],"origin":"Platform / user-generated","regions":["Not stated"],"languages":[],"sizeBand":"","count":"1M+ lifetime installs","volume":"1M+ lifetime app installs","years":"2019–2026","yearsOfHistory":"7","growing":"","systems":[],"contents":["VO2max","Resting heart rate","Heart-rate variability","Steps","Sleep"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Consumer wearable health app","tools":[],"span":"7 years","volumeHead":"1M lifetime installs","sizeTB":null,"fromYear":2019,"toYear":2026,"yearsNum":7.0,"tags":["wearables","health metrics","hrv","sleep","vo2max","fitness app","time series"],"shortTitle":"Wearable app health metrics"},{"ref":"MK-55064","title":"Software/SaaS company: 10+ year internal archive","headline":"Software/SaaS company: 10-19 years of email, chat, code, issue tracking, design files, database and HR system records","description":"Internal systems of a software/SaaS company in operation for 10 to 19 years: email, team chat, accounting, Git code repositories, cloud documents, an issue tracker, task management, design files, a Postgres database, e-signature and a HR system. The seller states it owns the data. Volumes are not stated.","category":"Company operations archive","sector":"Software & SaaS","subsector":"Software/SaaS company","dataTypes":["Code","Email","Chat & messaging","Documents","Tabular / structured"],"useCases":["Coding models","Enterprise workflow agents","Computer-use & tool-use agents","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"10-19","growing":"","systems":["Email (Gmail)","Team chat (Slack)","Accounting (QuickBooks)","Git (GitHub)","File storage (Google Drive)","Issue tracker (Jira)","Project management (Asana)","Design files (Figma)","Database (Postgres)","E-signature (DocuSign)","HR (Workday)"],"contents":["Code repositories","Issue tracker","Team chat","Email","Design files","Production database","Accounting","Signed documents","HR system records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Software/SaaS company","tools":[["slack","Slack"],["github","GitHub"],["jira","Jira"],["asana","Asana"],["gmail","Gmail"],["googledrive","Google Drive"],["quickbooks","QuickBooks"],["figma","Figma"],["postgresql","Postgres"],["hr","HR & payroll"],["sign","E-signature"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["saas","software","code","jira","slack","figma","postgres","company archive"],"shortTitle":"Decade-old SaaS company archive"},{"ref":"MK-55367","title":"Accounting workflows captured through a finance browser","headline":"Accounting data and workflows captured through a finance browser that companies use to access their systems of record","description":"Accounting data and workflows collected through a custom browser that client companies use to access their systems of record in order to integrate payments. The seller says it sees the full workflow and is helping these companies structure their data. Volumes, formats, time span and client rights have not yet been provided.","category":"Financial & transactional","sector":"Financial services & insurance","subsector":"Payments / accounting workflow software","dataTypes":["Tabular / structured","Screen recordings / UI traces"],"useCases":["Computer-use & tool-use agents","Enterprise workflow agents","Domain fine-tuning"],"origin":"Platform / user-generated","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"","systems":["Accounting","Browser"],"contents":["Accounting data from client systems of record","Accounting workflows","Payment-integration activity"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Payments / accounting workflow software","tools":[["accounting","Accounting"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["accounting","payments","browser","workflows","finance operations","systems of record"],"shortTitle":"Accounting browser workflows"},{"ref":"MK-55370","title":"US DTC subscription consumer brand: orders and ad-creative performance","headline":"Small US direct-to-consumer subscription brand: ~50K order/customer records plus Meta ad creatives with performance metrics","description":"Operating data from a small US direct-to-consumer subscription consumer packaged goods brand, about 4 years in operation. Includes ecommerce orders and customers (~50K records), hundreds of Meta ad creatives with performance metrics, email marketing flows and campaign performance, subscription billing and retention data, and creator/UGC content performance. History varies by source: ecommerce ~4 years, Meta ads reported as ~5 years (to be clarified), email and subscription ~3 years, UGC ~1 year. The owner wants to license anonymised ad performance and creative data; customer personal data is excluded without a separate agreement. Creator likeness and usage rights for UGC need checking.","category":"Consumer behaviour","sector":"Retail & e-commerce","subsector":"DTC subscription consumer packaged goods","dataTypes":["Tabular / structured","Image","Video","Text"],"useCases":["Domain fine-tuning","Forecasting & time series","Vision & perception"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"","count":"~50K order/customer records","volume":"~50K order/customer records; hundreds of Meta ad creatives with performance metrics; email flows and campaigns; subscription billing/retention; creator/UGC content performance","years":"","yearsOfHistory":"4","growing":"","systems":["Ecommerce","Ads (Meta)","Email marketing","Subscription billing"],"contents":["Ecommerce orders and customers (~50K)","Meta ad creatives with performance metrics","Email marketing flows and campaign performance","Subscription billing and retention","Creator/UGC content performance"],"formats":[],"delivery":"","labelled":"Partly","labelTypes":["Ad performance metrics"],"personalData":"Low","deidentification":"Anonymised; customer personal data excluded unless separately agreed","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"DTC subscription consumer packaged goods","tools":[["accounting","Accounting"],["analytics","Analytics & marketing"]],"span":"4 years","volumeHead":"50K order/customer records","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":4.0,"tags":["dtc","subscription","cpg","ad creatives","meta ads","email marketing","retention","ugc"],"shortTitle":"DTC subscription ad creatives"},{"ref":"MK-55909","title":"Residential real-estate brokerages: transactions and communications","headline":"US residential real-estate brokerages: transactions plus hundreds of thousands of calls, texts and emails a year","description":"Data from US residential real-estate firms introduced by an adviser who works with them: transaction records and the associated client calls, texts and emails. Firms handle around 300 transactions a year, some over 1,000, and generate hundreds of thousands of calls, texts and emails per year. Volumes in bytes, formats and time span have not yet been provided.","category":"Company operations archive","sector":"Construction & real estate","subsector":"Residential real-estate brokerage","dataTypes":["Email","Chat & messaging","Tabular / structured"],"useCases":["Enterprise workflow agents","Domain fine-tuning","SFT / instruction tuning"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"","count":"~300-1,000+ transactions per firm per year","volume":"~300-1,000+ transactions per firm per year; hundreds of thousands of calls, texts and emails per year","years":"","yearsOfHistory":"","growing":"Yes","systems":["Email","Telephony","CRM"],"contents":["Real-estate transaction records","Client call records","Text messages","Emails"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"Via authorised intermediary","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Residential real-estate brokerage","tools":[["email","Email"],["crm","CRM"],["calls","Calls"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["real estate","brokerage","transactions","calls","texts","email","residential"],"shortTitle":"Real-estate brokerage records"},{"ref":"MK-56113","title":"Direct-to-consumer online education business: 11+ years of CRM data across support, marketing and sales","headline":"One connected CRM linking ~2.4M customer support emails, ~40K funnel split tests with outcomes, ~43M sales emails and SMS, and the purchase history of ~3.8M customers","description":"The full CRM of a large direct-to-consumer online education and information-products business, built in-house and used across sales, marketing and customer support from 2014 to 2026. Customer support covers about 2.4M inbound customer emails and about 850K staff replies, alongside staff notes and logged calls. Marketing covers about 41K funnel split tests, 18K email and 5.7K SMS split tests across roughly 3,000 sales funnels, each with saved variants and results, plus tens of thousands of tracked paid-social ads with spend and, for several thousand, the ad creative. Sales covers about 34M sales emails and 10M SMS sent, about 455M logged page visits, 22M checkout visits and 1.2M purchases including recurring payments and refunds. Because everything sits in one database, a support question, a campaign, a funnel visit and a purchase can be linked to the same customer, which supports grounded support-agent tasks, marketing experimentation with measured outcomes, and conversion modelling. Export size and per-type volumes are being confirmed with the seller; customer and staff personal data will be de-identified before delivery and payment-card data excluded. The CRM's source code is not part of the offer.","category":"Company operations archive","sector":"Education","subsector":"Online education and information products (direct to consumer)","dataTypes":["Email","Customer support","CRM records","SMS","Web analytics / clickstream","Transactions","Ad creative and spend","Notes"],"useCases":["Customer-support agents","Marketing and copywriting","A/B test outcome prediction","Conversion modelling","Agent environments","Evals","Domain fine-tuning"],"origin":"Company internal records","regions":["North America"],"languages":["English"],"sizeBand":"1-10 TB","count":"~6.6B rows across 2,200+ tables","volume":"~1.4 TB CRM; ~3.8M customer profiles; ~2.4M support emails in and ~850K staff replies; ~41K funnel, 18K email and 5.7K SMS split tests across ~3K funnels; ~34M sales emails and ~10M SMS sent; ~455M page visits, ~22M checkout visits, ~1.2M purchases; ~25K tracked paid-social ads; ~220K staff notes; ~130K logged calls","years":"2014–2026","yearsOfHistory":"11+","growing":"Yes","systems":["Custom-built all-in-one CRM (MySQL)"],"contents":["Customer support threads with staff replies","Funnel, email and SMS split tests with results","Sales email and SMS campaigns","Page-visit, checkout and purchase events","Recurring payments and refunds","Paid-social ads with spend and creative","Staff notes and call logs"],"formats":["MySQL database exports"],"delivery":"Table exports (CSV/Parquet or SQL dump), de-identified, scope to be agreed","labelled":"Partly","labelTypes":["Split-test variants and results","Purchase and refund outcomes linked to campaigns and support contacts"],"personalData":"High","deidentification":"Planned before delivery: customer and staff identities removed, payment-card data excluded","rights":"To be confirmed","sample":"On request","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Online education and information products","tools":[["crm","CRM"]],"span":"12 years","volumeHead":"6.6B rows across 2,200+ tables","sizeTB":1.4,"fromYear":2014,"toYear":2026,"yearsNum":12,"tags":["crm","customer support","email marketing","sms marketing","a/b tests","split tests","sales funnels","conversion","e-commerce","online education","info products","ad spend","dtc"],"shortTitle":"DTC online education business CRM"},{"ref":"MK-57076","title":"Tech-candidate profiles with interview outcomes (India, US)","headline":"Tech-candidate profiles linked to hiring activity and interview outcomes (~15M entries, India and US)","description":"Records from a technical recruitment marketplace with over a million registered candidates, mostly in India and some in the US. Candidate profiles (resumes and salary information) are linked to hiring activity and detailed interview notes, including whether each candidate passed, all captured in the owner's own applicant-tracking tools. Contact details in resumes can be scrubbed before delivery.","category":"Workforce & labour market","sector":"Staffing & HR","subsector":"Technical recruitment marketplace","dataTypes":["Tabular / structured","Documents","Text"],"useCases":["Domain fine-tuning","Evaluation / benchmarks","Forecasting & time series","Enterprise workflow agents"],"origin":"Platform / user-generated","regions":["South Asia","North America"],"languages":["English"],"sizeBand":"","count":"~15M entries","volume":"~15M entries (1M+ candidates)","years":"2018–2026","yearsOfHistory":"8","growing":"Yes","systems":["Database","ATS"],"contents":["Candidate resumes","Salary information","Hiring activity","Interview notes","Interview pass/fail outcomes"],"formats":["CSV","Database dump"],"delivery":"Cloud bucket (S3)","labelled":"Yes","labelTypes":["Interview outcomes"],"personalData":"High","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"On request / under NDA","readiness":"Within 2 weeks","stage":"Summary available","scale":"","employees":"","industry":"Technical recruitment marketplace","tools":[["database","Database"],["hr","HR & payroll"]],"span":"8 years","volumeHead":"15M entries","sizeTB":null,"fromYear":2018,"toYear":2026,"yearsNum":8.0,"tags":["recruitment","resumes","interviews","hiring","candidates","salary","india","tech talent"],"shortTitle":"Tech candidate interview outcomes"},{"ref":"MK-57395","title":"Indian online-course app platform: teaching video, STEM Q&A and support chats","headline":"Indian app builder for online educators: tens of millions of teaching videos, millions of STEM Q&A, email and WhatsApp support groups over 7 years","description":"Seven years of data from an Indian company that builds apps for educators who sell online courses. The seller describes tens of millions of videos of teachers teaching, course content and audio in several languages, millions of STEM questions and answers, employee email, and thousands of WhatsApp support groups. Total size is stated as ~20 PB, held in cloud object storage. Course content may be owned by the educators who use the platform rather than the platform itself, so rights need confirming. Support groups contain end-user personal data and may include students who are minors.","category":"Video & imagery","sector":"Education","subsector":"App platform for online-course educators","dataTypes":["Video","Audio / speech","Documents","Text","Email","Chat & messaging"],"useCases":["Vision & perception","Speech & audio models","STEM reasoning","Domain fine-tuning","Retrieval / RAG"],"origin":"Platform / user-generated","regions":["South Asia"],"languages":["Multiple languages (not specified)"],"sizeBand":"100 TB+","count":"Tens of millions of videos","volume":"~20 PB stated; tens of millions of teaching videos; millions of STEM questions and answers; thousands of WhatsApp support groups; employee email","years":"","yearsOfHistory":"7","growing":"","systems":["Cloud storage (S3)","Email (Zoho)","WhatsApp"],"contents":["Teaching videos","Course content and audio","STEM questions and answers","Employee email","WhatsApp support group conversations","PDFs and documents"],"formats":["Video","Audio","PDF","Documents"],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"To be confirmed","sample":"","readiness":"1-2 weeks","stage":"Summary available","scale":"Mid-size company","employees":"51–200","industry":"App platform","tools":[["whatsapp","WhatsApp"],["email","Email"],["hr","HR & payroll"],["files","File storage"]],"span":"7 years","volumeHead":"Tens of millions of videos","sizeTB":20000.0,"fromYear":null,"toYear":null,"yearsNum":7.0,"tags":["edtech","online courses","teaching video","stem","q&a","whatsapp","india","multilingual"],"shortTitle":"Edtech lecture video archive"},{"ref":"MK-57875","title":"US logistics order-automation SaaS: orders, driver, scan and drop-off data","headline":"5 years of order, driver, location, scan and drop-off data from a US logistics SaaS working with ~120K drivers","description":"Data from a US logistics SaaS that automates orders with third-party delivery platforms, covering 5 years and about 200 GB. Includes hundreds of thousands of orders with driver, location, scan and drop-off data, across ~120K US drivers. Driver locations and customer addresses are present and need de-identification.","category":"Industrial & sensor","sector":"Logistics, distribution & wholesale","subsector":"Delivery order-automation SaaS","dataTypes":["Tabular / structured","Geospatial"],"useCases":["Forecasting & time series","Enterprise workflow agents","Domain fine-tuning"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"100 GB-1 TB","count":"~120K drivers","volume":"~200 GB; hundreds of thousands of orders; ~120K US drivers","years":"","yearsOfHistory":"5","growing":"","systems":[],"contents":["Orders","Driver records","Location data","Scan events","Drop-off records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"Needed (driver locations and customer addresses)","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Delivery order-automation SaaS","tools":[],"span":"5 years","volumeHead":"120K drivers","sizeTB":0.2,"fromYear":null,"toYear":null,"yearsNum":5.0,"tags":["logistics","last mile","delivery","drivers","scans","location data","orders"],"shortTitle":"US last-mile driver scan data"},{"ref":"MK-58758","title":"Small ADU builder: operations, calls and CAD","headline":"Small ADU construction company: 5-9 years of email, CRM, accounting, calls, field operations, CAD and contracts","description":"Records of a small company building accessory dwelling units (ADUs), in operation for 5 to 9 years: email, CRM, accounting, calls, field operations data, CAD files and contracts. Volumes are not stated.","category":"Company operations archive","sector":"Construction & real estate","subsector":"ADU construction","dataTypes":["Email","Tabular / structured","Audio / speech","CAD / 3D","Documents"],"useCases":["Enterprise workflow agents","CAD and design models","Domain fine-tuning"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"5-9","growing":"","systems":["Email","CRM","Accounting","Telephony","Field operations","CAD"],"contents":["Email","CRM records","Accounting","Calls","Field operations data","CAD files","Contracts"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"ADU construction","tools":[["email","Email"],["crm","CRM"],["calls","Calls"],["accounting","Accounting"],["design","Design files"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["construction","adu","cad","field operations","crm","calls"],"shortTitle":"ADU builder calls and CAD"},{"ref":"MK-59152","title":"Restaurant back-of-house data: 45M+ invoice lines linked to costed recipes and menu sales","headline":"Restaurant back-of-house data: 45M+ invoice lines linked to 815K recipes and 2.3M menu sales","description":"Operating data from ~395 US restaurant and foodservice operators using one back-of-house software product since 2017: 45M+ purchase invoice line items (distributor, item, pack size, quantity, unit price; ~10% from live EDI feeds of major US foodservice distributors), ~815K recipes and sub-recipes with quantities and yields (~331K fully costed against real invoice prices), ~2.3M recipe sales records, inventory counts and variance, and nutrition, allergen, food-safety and lot-tracking records. It links purchase cost to recipe to menu performance; operator names and negotiated prices are removed before delivery.","category":"Financial & transactional","sector":"Hospitality & food","subsector":"Restaurant back-of-house software","dataTypes":["Tabular / structured"],"useCases":["Forecasting & time series","Domain fine-tuning","Enterprise workflow agents","Evaluation / benchmarks"],"origin":"Platform / user-generated","regions":["North America"],"languages":["English"],"sizeBand":"","count":"45M+ invoice lines","volume":"45M+ invoice line items; ~815K recipes; ~2.3M recipe sales records","years":"2017–2026","yearsOfHistory":"9","growing":"Yes","systems":["Database"],"contents":["Invoice line items with unit prices","EDI distributor feeds","Costed recipes and sub-recipes","Recipe sales","Inventory counts and variance","Allergen and nutrition records"],"formats":["CSV","Parquet","JSON"],"delivery":"Cloud bucket (S3); Download; API","labelled":"Yes","labelTypes":["Recipe costing"],"personalData":"Low","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Within 2 weeks","stage":"Summary available","scale":"","employees":"","industry":"Restaurant back-of-house software","tools":[["database","Database"]],"span":"9 years","volumeHead":"45M invoice lines","sizeTB":null,"fromYear":2017,"toYear":2026,"yearsNum":9.0,"tags":["restaurants","foodservice","invoices","recipes","food cost","menu","procurement","edi"],"shortTitle":"Restaurant invoices and recipes"},{"ref":"MK-60172","title":"Enterprise back-office screen recordings with decision interviews; retail store operations","headline":"Enterprise back-office screen recordings with AI-led decision interviews; plus fuel, QSR and convenience-store ops data","description":"Two assets. (1) Screen recordings of real enterprise workflows (requisition-to-PO, inventory replenishment, RFQs and quotes, invoice processing, cash application, support tickets) in tools such as SAP, NetSuite, Concur, Excel and email, across procurement, supply chain, utilities, pharma, finance, logistics and engineering/construction, each paired with an AI-led interview in which the worker explains their decisions; these would be brokered with participating customers' permission. (2) Separately, store-level operating data (payroll, inventory and possibly POS, purchasing and accounting) from owner-operated fuel stations, quick-service restaurants and convenience stores.","category":"Expert workflows","sector":"AI & data","subsector":"Enterprise workflow capture; retail store operations","dataTypes":["Screen recordings / UI traces","Video","Text","Tabular / structured"],"useCases":["Computer-use & tool-use agents","Enterprise workflow agents","SFT / instruction tuning","Evaluation / benchmarks","Forecasting & time series"],"origin":"Mixed","regions":["North America"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"Yes","systems":["ERP","Expense management","Spreadsheets","Email","Payroll","POS","Accounting"],"contents":["Requisition-to-PO screen recordings","Inventory replenishment workflows","RFQ and quote handling","Invoice processing","Cash application","Support-ticket handling","AI interview transcripts explaining decisions","Store payroll records","Store inventory records","POS and purchasing data (possible)"],"formats":["Screen recordings","Interview transcripts","Store records"],"delivery":"","labelled":"Partly","labelTypes":["Decision explanations from worker interviews"],"personalData":"High","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Enterprise workflow capture","tools":[["sheet","Excel"],["erp","NetSuite"],["sap","SAP"],["email","Email"],["accounting","Accounting"],["shop","Commerce"],["hr","HR & payroll"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["screen recordings","enterprise workflows","procurement","invoice processing","sap","netsuite","decision interviews","retail","convenience stores","payroll"],"shortTitle":"Back-office screen recordings"},{"ref":"MK-61278","title":"Egocentric video of road-construction work in India (to order)","headline":"First-person video of real road-construction tasks, collected to order through 20+ contractors with thousands of workers","description":"Proposed collection of egocentric video of workers doing real physical tasks on road-construction sites in India, through a network of 20+ contractors (thousands of workers, scalable to hundreds of contractors).","category":"Physical-world video","sector":"Construction & real estate","subsector":"Road construction","dataTypes":["Video"],"useCases":["Robotics / embodied AI","World models"],"origin":"Purpose-collected (consented capture)","regions":["South Asia"],"languages":[],"sizeBand":"","count":"","volume":"To order","years":"Not yet collected","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Collected to order","stage":"Collected to order","scale":"","employees":"","industry":"Road construction","tools":[],"span":"","volumeHead":"To order","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["egocentric","construction","video","physical ai"],"shortTitle":"Indian road-construction video"},{"ref":"MK-61489","title":"Game-engine startup: team chat, pre-AI animation pipeline and user prompts","headline":"Game-engine startup archive: team chat, pre-AI Blender animation pipeline, code, error tickets and user prompts","description":"A small US startup that spent three years building a game engine used by about 15,000 developers. The archive holds about 30,000 team-chat messages from six people building the engine and an automated Blender-based animation and hair-simulation generator built before video models, tens of thousands of lines of pre-AI code, user-reported error tickets with the submissions that triggered them, and thousands of user first prompts.","category":"Code & software","sector":"Software & internet","subsector":"Game engine startup","dataTypes":["Chat & messaging","Code","Documents","Text"],"useCases":["Coding models","Pre-training","SFT / instruction tuning","Evaluation / benchmarks"],"origin":"First-party company records","regions":["North America"],"languages":["English"],"sizeBand":"100 GB-1 TB","count":"~30K team messages","volume":"~1 TB","years":"2026","yearsOfHistory":"4","growing":"No","systems":["Team chat","File storage","Git","Issue tracker"],"contents":["Team chat (~30K messages)","Game-engine source code","Blender animation pipeline","Hair-simulation pipeline","Error tickets with triggering submissions","User first prompts","Planning documents and presentations"],"formats":[],"delivery":"","labelled":"No","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Within 2 weeks","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Game engine startup","tools":[["chat","Team chat"],["tickets","Tickets"],["git","Git"],["files","File storage"]],"span":"4 years","volumeHead":"30K team messages","sizeTB":1.0,"fromYear":null,"toYear":2026,"yearsNum":4.0,"tags":["game engine","team chat","code","blender","animation","error tickets","user prompts","startup archive"],"shortTitle":"Game-engine startup archive"},{"ref":"MK-62058","title":"Live news and culture knowledge graph","headline":"Time-stamped knowledge graph of news and culture entities, events and relationships, refreshed every 30 minutes","description":"Structured knowledge graph built from the live web, covering entities, events and relationships in news and culture and tracking how they change over time. Refreshed every 30 minutes and delivered as clean, time-stamped structured data via API or files, cut to order. Pitched for training, real-time grounding and temporal-reasoning evaluations. Size, history depth and rights to the underlying web sources are not yet stated. A sample is offered.","category":"Text & documents","sector":"AI & data","subsector":"Web-derived knowledge graph","dataTypes":["Tabular / structured"],"useCases":["Retrieval / RAG","Grounding","Temporal-reasoning evaluation"],"origin":"Public / scraped sources","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"Yes","systems":[],"contents":["Entities","Events","Relationships","Change history over time"],"formats":["API or files"],"delivery":"API or files","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"On request","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Web-derived knowledge graph","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["knowledge graph","news","culture","entities","events","temporal reasoning","grounding"],"shortTitle":"Live news and culture graph"},{"ref":"MK-62686","title":"International corporate law firm: 35-year archive","headline":"Small international corporate law firm: 35 years of legal documents (~800 GB excluding email), de-identified by the firm","description":"Document archive of a small international corporate law firm covering 35 years, about 800 GB excluding emails and messages, roughly 70% in English. Contents are client legal matters and contracts. The firm says it will de-identify all data before transfer using specialised legal software. Privilege and client-confidentiality conditions are to be confirmed.","category":"Text & documents","sector":"Legal","subsector":"International corporate law firm","dataTypes":["Documents","Text"],"useCases":["Domain fine-tuning","Retrieval / RAG","Enterprise workflow agents"],"origin":"First-party company records","regions":["Not stated"],"languages":["English (~70%)","other languages"],"sizeBand":"100 GB-1 TB","count":"~800 GB","volume":"~800 GB excluding emails and messages; 35 years","years":"35 years","yearsOfHistory":"35","growing":"","systems":[],"contents":["Client legal matters","Contracts","Corporate legal documents"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"Owner to de-identify before transfer","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Timing to confirm","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"International corporate law firm","tools":[],"span":"35 years","volumeHead":"800 GB","sizeTB":0.8,"fromYear":null,"toYear":null,"yearsNum":35.0,"tags":["legal","law firm","corporate law","contracts","long history","documents"],"shortTitle":"35-year corporate law firm archive"},{"ref":"MK-63654","title":"Architecture firm: photos, drone video, CAD and project data","headline":"US architecture firm: ~15 TB of project photos, drone video, CAD, project-management and invoice data","description":"Project archive of a US architecture firm: site and project photographs, drone video, CAD files and project-management data, and the firm is also open to licensing all of its invoice data. The archive is stated at about 15 TB. The same adviser also represents engineering, finance and energy companies whose data is not yet described. A filled inventory is pending.","category":"Company operations archive","sector":"Construction & real estate","subsector":"Architecture firm","dataTypes":["Image","Video","3D / CAD","Tabular / structured"],"useCases":["Vision & perception","World models","Domain fine-tuning","Enterprise workflow agents"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"10-100 TB","count":"~15 TB","volume":"~15 TB","years":"","yearsOfHistory":"","growing":"","systems":["Design files","Project management","Accounting"],"contents":["Project photographs","Drone video","CAD files","Project-management records","Invoice data"],"formats":["Photos","Drone video","CAD","Project-management exports"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Architecture firm","tools":[["tickets","Tickets"],["accounting","Accounting"],["design","Design files"]],"span":"","volumeHead":"15 TB","sizeTB":15.0,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["architecture","cad","drone video","photos","project management","invoices","construction","aec"],"shortTitle":"Architecture firm projects"},{"ref":"MK-64780","title":"Small construction company: email and CAD files","headline":"Small construction company: 5-9 years of email plus AutoCAD and Revit CAD files","description":"Records of a small construction company in operation for 5 to 9 years: business email plus AutoCAD and Revit CAD files. The seller states it owns the data. Volumes are not stated.","category":"Company operations archive","sector":"Construction & real estate","subsector":"Construction","dataTypes":["Email","CAD / 3D"],"useCases":["CAD and design models","Domain fine-tuning","Enterprise workflow agents"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"5-9","growing":"","systems":["Email (Gmail)","CAD (AutoCAD, Revit)"],"contents":["Business email","AutoCAD drawings","Revit models"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Construction","tools":[["gmail","Gmail"],["autocad","AutoCAD"],["design","Revit"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["construction","cad","autocad","revit","email"],"shortTitle":"Small builder email and CAD"},{"ref":"MK-65917","title":"Portfolio of five US mid-market businesses","headline":"Five US mid-market businesses: acoustics projects (~2.4 TB), plumbing CRM, funeral roll-up diligence, boat builder, finance firm","description":"Data from five US businesses each with roughly $5M-$20M revenue: (1) an acoustics manufacturer and contractor with organised data for hundreds of projects over 20+ years (~2.4 TB); (2) vertically integrated plumbing companies (two consumer brands plus a waterproofing manufacturer/distributor) with nearly 200K CRM contacts and about 80 cumulative years of operation; (3) a funeral-services roll-up with 10+ acquisitions, holding acquisition documents, diligence, SOPs, quality-of-earnings reports, legal and operational files, management communications, PE financial models and capital-raising documents, plus a related 6-year-old funeral business with ~80 GB of email and ~11K CRM contacts; (4) a legacy PE-backed small-craft boat builder; and (5) a long-standing finance, consulting and banking firm (capital raising, strategy, advisory and broker-dealer arms) with thousands of projects. Sources include email, CRM, accounting software and on-premise project-file servers.","category":"Company operations archive","sector":"Other","subsector":"Multi-company portfolio (acoustics, plumbing, funeral services, boat building, finance)","dataTypes":["Email","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Pre-training","Retrieval / RAG"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"1-10 TB","count":"5 companies; ~2.5 TB stated so far","volume":"Acoustics 2.4 TB; plumbing ~200K CRM contacts; funeral business 80 GB email + 11K contacts","years":"Up to 20+ years per company","yearsOfHistory":"20","growing":"","systems":["Email","CRM","Accounting","File storage"],"contents":["Acoustics project files (hundreds of projects)","Plumbing CRM contacts (~200K)","Acquisition and diligence documents","SOPs","Quality-of-earnings reports","PE financial models","Capital-raising documents","Management communications","Email correspondence (~80 GB, funeral business)","Finance/consulting project records (thousands of projects)"],"formats":["Email","CRM exports","Accounting data","Project files"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Via authorised intermediary","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Multi-company portfolio","tools":[["email","Email"],["crm","CRM"],["accounting","Accounting"],["files","File storage"]],"span":"20 years","volumeHead":"5 companies","sizeTB":2.5,"fromYear":null,"toYear":null,"yearsNum":20.0,"tags":["multi-company","acoustics","plumbing","funeral services","boat building","private equity","diligence","crm","email","financial models"],"shortTitle":"Five mid-market businesses"},{"ref":"MK-66194","title":"1:1 professional coaching sessions: 10K+ hours of audio and video","headline":"10K+ hours of recorded 1:1 coaching sessions (career, consulting, finance, admissions)","description":"Recorded one-to-one coaching sessions (audio and video) from a coaching marketplace with thousands of coaches, covering hundreds of topics across consulting, finance, admissions, AI and career coaching, mostly in English. Pitched for voice and conversational model training and coaching-domain context. The owner will scrub personal content; coach and client consent for third-party AI training is still to be confirmed.","category":"Speech & conversation","sector":"Education","subsector":"Coaching marketplace","dataTypes":["Audio / speech","Video"],"useCases":["Speech & audio models","SFT / instruction tuning","Domain fine-tuning"],"origin":"Platform / user-generated","regions":["Global"],"languages":["English"],"sizeBand":"","count":"10K+ hours","volume":"10K+ hours","years":"2025–2026","yearsOfHistory":"1","growing":"Yes","systems":[],"contents":["1:1 coaching sessions","Career and admissions coaching","Consulting and finance case coaching"],"formats":["MP4"],"delivery":"","labelled":"Partly","labelTypes":[],"personalData":"High","deidentification":"Planned / on request","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"Within 2 weeks","stage":"Summary available","scale":"","employees":"","industry":"Coaching marketplace","tools":[],"span":"1 year","volumeHead":"10K hours","sizeTB":null,"fromYear":2025,"toYear":2026,"yearsNum":1.0,"tags":["coaching","conversations","voice","speech","career","admissions","consulting"],"shortTitle":"Professional coaching sessions"},{"ref":"MK-67370","title":"Financial-services company: 10-19 years of email, code and databases","headline":"Small financial-services company offering 10-19 years of email, Git repositories and PostgreSQL/MongoDB databases","description":"Operating data of a small financial-services company with hundreds of customers, covering 10-19 years: Gmail, Git repositories (GitHub and Bitbucket) and PostgreSQL/MongoDB databases. Volumes not yet stated.","category":"Company operations archive","sector":"Financial services & insurance","subsector":"","dataTypes":["Email","Code","Tabular / structured"],"useCases":["Coding models","Enterprise workflow agents","Domain fine-tuning"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"Hundreds of customers","years":"","yearsOfHistory":"10-19","growing":"","systems":["Email (Gmail)","Git (GitHub","Bitbucket)","Database (PostgreSQL","MongoDB)"],"contents":["Email","Code repositories","PostgreSQL and MongoDB databases"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Financial services & insurance","tools":[["github","GitHub"],["gmail","Gmail"],["database","Database"]],"span":"","volumeHead":"Hundreds of customers","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["financial services","code","github","bitbucket","postgresql","mongodb","email"],"shortTitle":"Financial firm code and databases"},{"ref":"MK-67808","title":"Unpublished English STEM long-form works: ~20K documents","headline":"~20K unpublished English long-form STEM works (~260M words) across engineering, biology, medicine, chemistry, physics and maths","description":"A collection of ~20K unpublished English long-form research works in engineering, biology, medical sciences, chemistry, physics and mathematics, averaging 12-13K words each (~260M words in total), as PDF and DOC files. Samples seen are thesis and project-style write-ups with author names removed. The seller claims control of commercial licensing rights; some works are tied to blended-learning programmes, and the authorship and rights chain is still being confirmed. The seller says the works span 20+ years.","category":"Text & documents","sector":"Education","subsector":"Unpublished STEM research and project writing","dataTypes":["Documents","Text"],"useCases":["Domain fine-tuning","Reasoning & science","Retrieval / RAG"],"origin":"","regions":["Not stated"],"languages":["English"],"sizeBand":"","count":"~20K research works","volume":"~20K works; avg 12-13K words each; ~260M words","years":"20+ years, revised to 2025/26","yearsOfHistory":"20+","growing":"","systems":[],"contents":["Long-form works in engineering, biology, medical sciences, chemistry, physics and mathematics","Methods, results and discussion sections"],"formats":["PDF","DOC"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"Author names removed in samples","rights":"To be confirmed","sample":"Available","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Unpublished STEM research and project writing","tools":[],"span":"","volumeHead":"20K research works","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["stem","research writing","long-form text","engineering","biology","chemistry","physics","mathematics"],"shortTitle":"Unpublished STEM long-form works"},{"ref":"MK-67825","title":"Mid-size software/SaaS company: 5-9 year company archive","headline":"Mid-size software/SaaS company: 5-9 years of email, chat, CRM, support tickets, code, calls, wiki and database","description":"Company data from a mid-size software/SaaS business with thousands of customers and 5-9 years of history. Systems listed: email, team chat, CRM, accounting, customer support, code, calls, file storage, an internal wiki and a database. Submitted by an operations lead rather than the owner. Volumes are not yet stated.","category":"Company operations archive","sector":"Software & SaaS","subsector":"Software / SaaS","dataTypes":["Email","Chat & messaging","Code","Audio / speech","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Coding models","Speech & audio models","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"5-9 years","yearsOfHistory":"5-9","growing":"","systems":["Email","Team chat","CRM","Accounting","Support desk","Git","Telephony","File storage","Wiki","Database"],"contents":["Email","Team chat","CRM","Accounting","Support tickets","Code","Call recordings","Files","Wiki","Database"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"Mid-size company","employees":"51–200","industry":"Software / SaaS","tools":[["chat","Team chat"],["git","Git"],["email","Email"],["crm","CRM"],["wiki","Docs & wiki"],["calls","Calls"],["database","Database"],["accounting","Accounting"],["files","File storage"]],"span":"9 years","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["saas","support tickets","code","crm","calls","wiki"],"shortTitle":"Mid-size SaaS support and code"},{"ref":"MK-68056","title":"Publisher content licensing (2,000+ publishers)","headline":"Publisher content and data licensed through an intermediary representing 2,000+ publishers","description":"Content and data licensed on behalf of a network of more than 2,000 publishers worldwide, offered through a licensing intermediary acting for those publishers. Content types, volumes and the rights held per publisher have not yet been provided.","category":"Text & documents","sector":"Marketing, media & publishing","subsector":"Publisher content licensing","dataTypes":["Text","Documents"],"useCases":["Pre-training","Retrieval / RAG","Domain fine-tuning"],"origin":"Licensed from third parties","regions":["Global"],"languages":[],"sizeBand":"","count":"2,000+ publishers","volume":"2,000+ publishers","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":["Publisher content","Publisher data","2,000+ publisher network"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Via authorised intermediary","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Publisher content licensing","tools":[],"span":"","volumeHead":"2K publishers","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["publishing","content licensing","publishers","text","media","pre-training"],"shortTitle":"Publisher content licensing"},{"ref":"MK-68740","title":"Logistics company: 20+ years of email, CRM, accounting and field operations","headline":"Small logistics company offering 20+ years of Exchange email, CRM, QuickBooks accounting, field-operations records and contracts","description":"Operating data of a small logistics company covering 20+ years: Exchange email, CRM, QuickBooks accounting, field-operations records and contracts. Volumes not yet stated.","category":"Company operations archive","sector":"Logistics, distribution & wholesale","subsector":"Logistics services","dataTypes":["Email","Tabular / structured","Documents"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Forecasting & time series"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"20+","growing":"","systems":["Email (Exchange)","CRM","Accounting (QuickBooks)","Field operations"],"contents":["Email","CRM records","Accounting records","Field-operations records","Contracts"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Logistics services","tools":[["quickbooks","QuickBooks"],["email","Email"],["crm","CRM"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["logistics","field operations","exchange","crm","quickbooks","contracts"],"shortTitle":"20-year logistics firm field ops"},{"ref":"MK-73577","title":"US real-estate virtual staging service: 9-year operations archive","headline":"US virtual staging service for real estate agents: ~2 TB across email, chat, CRM, support, calls, files and cloud storage","description":"Nine years of operating data from a small US virtual staging business serving real estate agents, with thousands of customers. About 2 TB across email, Slack, HighLevel CRM, Intercom support, Zoom, Google Drive and AWS. Export expected in 1-2 weeks.","category":"Company operations archive","sector":"Real estate & construction","subsector":"Virtual staging for real estate agents","dataTypes":["Email","Chat & messaging","Image","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Vision & perception","Domain fine-tuning"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"1-10 TB","count":"~2 TB","volume":"~2 TB over 9 years; thousands of customers","years":"","yearsOfHistory":"9","growing":"","systems":["Email (Gmail)","Team chat (Slack)","CRM (HighLevel)","Customer support (Intercom)","Video calls (Zoom)","File storage (Google Drive)","Cloud storage (AWS)"],"contents":["Email","Team chat","CRM records","Support conversations","Zoom meetings","Drive files","Cloud-stored files"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Export 1-2 weeks","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Virtual staging","tools":[["slack","Slack"],["intercom","Intercom"],["gmail","Gmail"],["googledrive","Google Drive"],["zoom","Zoom"],["crm","CRM"]],"span":"9 years","volumeHead":"2 TB","sizeTB":2.0,"fromYear":null,"toYear":null,"yearsNum":9.0,"tags":["virtual staging","real estate","intercom","crm","slack","images"],"shortTitle":"Virtual staging business archive"},{"ref":"MK-74258","title":"US veterinary business: in-house practice software and code","headline":"Small US veterinary business: in-house practice-management system, code, database and email from 15 inboxes, ~20 GB","description":"Data of a small US veterinary business that built its own software, including a practice-information management system (PIMS). Includes the PIMS and its database, Git code repositories, cloud documents and email from 15 inboxes, about 20 GB over 3 years. The PIMS holds pet-owner and client records.","category":"Company operations archive","sector":"Healthcare & life sciences","subsector":"Veterinary practice with in-house software","dataTypes":["Code","Tabular / structured","Email","Documents"],"useCases":["Coding models","Enterprise workflow agents","Domain fine-tuning"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"10-100 GB","count":"15 email inboxes","volume":"~20 GB; 15 inboxes","years":"","yearsOfHistory":"3","growing":"","systems":["Email (Gmail)","Git (GitHub)","File storage (Google Drive)","Database","Practice management (PIMS)"],"contents":["Custom practice-management system","Code repositories","Database","Email","Documents"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Export 1-2 weeks","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Veterinary practice","tools":[["github","GitHub"],["gmail","Gmail"],["googledrive","Google Drive"],["database","Database"]],"span":"3 years","volumeHead":"15 email inboxes","sizeTB":0.02,"fromYear":null,"toYear":null,"yearsNum":3.0,"tags":["veterinary","pims","practice management","code","pet care"],"shortTitle":"Vet practice in-house PIMS"},{"ref":"MK-75118","title":"STEM research works and professional material","headline":"20,000+ research works, mainly English-language STEM, plus specialist professional material","description":"More than 20,000 research works, primarily in English and mainly STEM, plus specialist professional and analytical material, offered as a content partnership. Delivered as documents; formats, years and rights per work have not yet been stated.","category":"Text & documents","sector":"Marketing, media & publishing","subsector":"Research and professional content","dataTypes":["Documents","Text"],"useCases":["Pre-training","Domain fine-tuning","Retrieval / RAG"],"origin":"","regions":["Europe"],"languages":["English"],"sizeBand":"","count":"20,000+ research works","volume":"20,000+ research works","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":["STEM research works","Specialist professional material","Analytical material"],"formats":["Documents"],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Research and professional content","tools":[],"span":"","volumeHead":"20K research works","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["research","stem","academic","documents","english","professional content"],"shortTitle":"STEM research works"},{"ref":"MK-75670","title":"US enterprise consulting firm: 10-year archive","headline":"US enterprise consulting firm: 10 years of client work, plans, SOPs and CRM history (~40 TB)","description":"Ten years of records from a US consulting firm serving enterprise companies: about 12,000 folders of consulting work, business development plans, SOPs and data in file storage, plus CRM history with invoices and deal-call records. About 40 TB, exportable within two weeks.","category":"Company operations archive","sector":"Professional services","subsector":"Enterprise consulting","dataTypes":["Documents","Email","Tabular / structured"],"useCases":["Enterprise workflow agents","Pre-training","Domain fine-tuning"],"origin":"First-party company records","regions":["North America"],"languages":["English"],"sizeBand":"10-100 TB","count":"~12,000 project folders","volume":"~40 TB","years":"2026","yearsOfHistory":"10","growing":"Yes","systems":["CRM","File storage","Email"],"contents":["Consulting deliverables","Business development plans","SOPs","Invoices","Deal-call records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Within 2 weeks","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Enterprise consulting","tools":[["email","Email"],["crm","CRM"],["files","File storage"]],"span":"10 years","volumeHead":"12K project folders","sizeTB":40.0,"fromYear":null,"toYear":2026,"yearsNum":10.0,"tags":["consulting","company archive","sops","business plans","crm","professional services","usa"],"shortTitle":"Enterprise consulting archive"},{"ref":"MK-75840","title":"Factory and construction-site recordings (to order)","headline":"Live recordings from factories and construction sites in Pakistan and China (to order)","description":"Physical-world data to be recorded live in factories and on construction sites in Pakistan and China. Recording equipment was about to be deployed at sites already arranged; volume, modalities and formats are not yet stated.","category":"Physical-world video","sector":"Manufacturing","subsector":"Factories and construction sites","dataTypes":["Video"],"useCases":["Robotics / embodied AI","World models"],"origin":"Purpose-collected (consented capture)","regions":["South Asia","East Asia"],"languages":[],"sizeBand":"","count":"","volume":"","years":"Not yet collected","yearsOfHistory":"","growing":"","systems":[],"contents":["Factory recordings","Construction-site recordings"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Collected to order","stage":"Collected to order","scale":"","employees":"","industry":"Factories and construction sites","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["physical ai","factory","construction","video","pakistan","china","robotics"],"shortTitle":"Factory and site recordings"},{"ref":"MK-76053","title":"Tender and bid-response SaaS: code, workspace data and database (1 year)","headline":"Early-stage tender and bid-response software company: GitHub code, Google Workspace and PostgreSQL database, ~10 GB","description":"Data from a young, founder-run software company building a tender and bid-response product. Includes GitHub code, Google Workspace data and a PostgreSQL database, about 10 GB covering one year.","category":"Code & software","sector":"Software & SaaS","subsector":"Tender and bid-response software","dataTypes":["Code","Email","Documents","Tabular / structured"],"useCases":["Coding models","Enterprise workflow agents"],"origin":"First-party company records","regions":["North America","Europe"],"languages":[],"sizeBand":"10-100 GB","count":"~10 GB","volume":"~10 GB across code, workspace and database","years":"","yearsOfHistory":"1","growing":"","systems":["GitHub","Google Workspace","Database (PostgreSQL)"],"contents":["Source code","Workspace email and documents","Product database"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Tender and bid-response software","tools":[["github","GitHub"],["database","Database"]],"span":"1 year","volumeHead":"10 GB","sizeTB":0.01,"fromYear":null,"toYear":null,"yearsNum":1.0,"tags":["saas","tenders","bid response","code","postgresql"],"shortTitle":"Tender-response SaaS codebase"},{"ref":"MK-76298","title":"UK workplace-wellbeing SaaS (in a sale process): 4 years of source code plus 6 years of engineering wikis","headline":"Source code and Confluence operational and architecture documentation of a UK workplace mental-health platform, offered through an asset sale; customer and health records excluded","description":"Source code (4 years) and internal Confluence wikis with operations and architecture documentation (6 years) of a UK workplace wellbeing SaaS platform being sold through an asset-sale process. Customer and health records are excluded; 156 GB across all systems, eligible subset being defined.","category":"Code & software","sector":"Software & SaaS","subsector":"Workplace wellbeing","dataTypes":["Code","Documents"],"useCases":["Coding models","Enterprise workflow agents"],"origin":"First-party company records","regions":["Europe"],"languages":[],"sizeBand":"","count":"","volume":"~50 MB code and docs; 156 GB across systems (eligible subset TBD)","years":"4-6 years","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Export about a month","stage":"Summary available","scale":"","employees":"","industry":"Workplace wellbeing","tools":[["confluence","Confluence"]],"span":"6 years","volumeHead":"50 MB code and docs","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["saas","source code","confluence","architecture docs"],"shortTitle":"Workplace-wellbeing SaaS code"},{"ref":"MK-80610","title":"Small software/SaaS company: 2-4 year company archive","headline":"Small software/SaaS company: email, team chat, CRM, code, call recordings, files, database, contracts and HR over 2-4 years","description":"Company data from a small software/SaaS business with 2-4 years of history. Systems listed: email, team chat and messaging apps, CRM, accounting, code repositories, video-call recordings, file storage, an internal wiki, a database, contracts and HR records. Volumes are not yet stated.","category":"Company operations archive","sector":"Software & SaaS","subsector":"Software / SaaS startup","dataTypes":["Email","Chat & messaging","Code","Audio / speech","Video","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Coding models","Speech & audio models","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"2-4 years","yearsOfHistory":"2-4","growing":"","systems":["Email (Gmail)","Team chat (Slack, Discord)","WhatsApp","CRM (HubSpot)","Accounting (QuickBooks)","GitHub","Video calls (Zoom, Meet)","File storage (Google Drive)","Wiki","Database","Contracts","HR"],"contents":["Email","Slack, Discord and WhatsApp chat","HubSpot CRM","QuickBooks accounting","GitHub code","Zoom/Meet recordings","Drive files","Wiki","Database","Contracts","HR records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Software / SaaS startup","tools":[["slack","Slack"],["discord","Discord"],["whatsapp","WhatsApp"],["github","GitHub"],["hubspot","HubSpot"],["gmail","Gmail"],["googledrive","Google Drive"],["quickbooks","QuickBooks"],["zoom","Zoom"],["wiki","Docs & wiki"],["database","Database"],["hr","HR & payroll"]],"span":"4 years","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["saas","startup","code","crm","call recordings","team chat"],"shortTitle":"Small SaaS startup full stack"},{"ref":"MK-80927","title":"Coding-agent traces from a free coding assistant (~2K, growing)","headline":"Full traces from a free coding agent: ~2K sessions so far and growing fast; can be stripped of user code","description":"Complete traces from a free, ad-funded coding agent: a couple of thousand sessions so far, growing fast. Traces can be stripped; no repository IP captured.","category":"Agent traces","sector":"AI & data","subsector":"Coding agent","dataTypes":["Agent trajectories","Code"],"useCases":["Coding models","Computer-use & tool-use agents"],"origin":"Platform / user-generated","regions":["Not stated"],"languages":[],"sizeBand":"","count":"~2K traces","volume":"~2K traces","years":"2026","yearsOfHistory":"0","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Coding agent","tools":[],"span":"","volumeHead":"2K traces","sizeTB":null,"fromYear":2026,"toYear":2026,"yearsNum":null,"tags":["coding agent","traces","tool use"],"shortTitle":"Coding-agent traces"},{"ref":"MK-81019","title":"Marketing agency: 5-9 year operational database","headline":"Database from a small marketing agency with tens of thousands of customers, covering 5-9 years","description":"Database export from a small marketing agency with a customer base in the tens of thousands, covering 5-9 years. The seller has listed the database only; tables, volumes and contents have not been described yet.","category":"Company operations archive","sector":"Marketing, media & publishing","subsector":"Marketing agency","dataTypes":["Tabular / structured"],"useCases":["Enterprise workflow agents","Domain fine-tuning"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"5-9","growing":"","systems":["Database"],"contents":["Operational database"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Marketing agency","tools":[["database","Database"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["marketing agency","database","structured data"],"shortTitle":"Marketing agency database"},{"ref":"MK-82209","title":"Wholesale distributor: orders, suppliers, pricing and logistics","headline":"Wholesale distributor (11+ years): email, orders, suppliers, pricing and logistics from its management system","description":"Records of a wholesale and distribution company operating for 11+ years: mostly email (the last 2-3 years are retained, as older mail is trimmed periodically) plus orders, suppliers, pricing and logistics data available as a full export from its proprietary management software. The company also runs in-house CNC machining, so CAD drawings for parts are probably available. Only an introduction has been received so far.","category":"Company operations archive","sector":"Logistics, distribution & wholesale","subsector":"Wholesale distributor with in-house CNC","dataTypes":["Email","Tabular / structured","3D / CAD"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Forecasting & time series","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"Email covers the last 2-3 years","years":"Company 11+ years","yearsOfHistory":"","growing":"Yes","systems":["Email","ERP","Design files"],"contents":["Email (last 2-3 years)","Orders","Supplier records","Pricing","Logistics records","CAD drawings for CNC parts (probable)"],"formats":["Full export from proprietary management software","Email"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Wholesale distributor","tools":[["email","Email"],["erp","ERP"],["design","Design files"]],"span":"","volumeHead":"Email covers the last 2-3 years","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["wholesale","distribution","orders","suppliers","pricing","logistics","email","cnc","cad"],"shortTitle":"Wholesale distributor records"},{"ref":"MK-82218","title":"Robotic cell-biology data generated to order: calcium imaging and morphology on human cells","headline":"Data generated to order on a robotic wet-lab platform: calcium pharmacology on human iPSC neurons, tumour-stroma calcium responses and label-free morphology across perturbations","description":"A robotic platform grows, treats, stains and images 96-well plates of primary human cells to produce datasets to order. Example packages: calcium pharmacology on human iPSC-derived neurons with ~40 modulators (single-cell traces, ~4 months); calcium responses in normal, tumour-adjacent and tumour stroma with ~30 agonists (~4 months, sample available); label-free morphology across 20 perturbations (~2 months). Also offers an automation-ready chemistry space of ~1M compounds. Rights assigned or licensed to the buyer.","category":"Healthcare & life sciences","sector":"Healthcare & life sciences","subsector":"Cell biology / wet lab","dataTypes":["Image","Tabular / structured"],"useCases":["Domain fine-tuning","World models"],"origin":"Purpose-collected (consented capture)","regions":["Europe"],"languages":[],"sizeBand":"","count":"","volume":"To order (96-well plate runs)","years":"2-4 months per package","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"Yes","labelTypes":[],"personalData":"None","deidentification":"Not needed","rights":"Owner-held","sample":"Available","readiness":"Collected to order","stage":"Collected to order","scale":"","employees":"","industry":"Cell biology / wet lab","tools":[],"span":"","volumeHead":"To order","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["biology","calcium imaging","ipsc","cell imaging","wet lab","to order"],"shortTitle":"Robotic cell-biology data"},{"ref":"MK-84003","title":"Construction company: 10-19 years of email, accounting, recordings, CAD and contracts","headline":"Construction company archive: Outlook email, QuickBooks, Teams recordings, SharePoint and OneDrive files, AutoCAD drawings, DocuSign contracts and HR over 10-19 years","description":"Operating data from a construction company with 10-19 years of history and a small number of large customers. The seller lists Outlook email, QuickBooks accounting, Teams recordings, SharePoint and OneDrive files, AutoCAD drawings, DocuSign contracts and HR records. Volumes have not been provided yet.","category":"Company operations archive","sector":"Construction & real estate","subsector":"Construction contractor","dataTypes":["Email","Tabular / structured","Video","Audio / speech","Documents","CAD / drawings"],"useCases":["Enterprise workflow agents","Engineering & CAD models","Speech & audio models","Domain fine-tuning"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"10-19","growing":"","systems":["Outlook","QuickBooks","Meeting recordings (Teams)","SharePoint","OneDrive","AutoCAD","DocuSign","HR"],"contents":["Email","Accounting records","Meeting recordings","Project files","AutoCAD drawings","Signed contracts","HR records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Timing to confirm","stage":"Summary available","scale":"Large organisation","employees":"200+","industry":"Construction contractor","tools":[["email","Outlook"],["files","SharePoint"],["files","OneDrive"],["quickbooks","QuickBooks"],["autocad","AutoCAD"],["git","Git"],["hr","HR & payroll"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["construction","cad","autocad","contracts","meeting recordings","accounting"],"shortTitle":"Contractor CAD and project archive"},{"ref":"MK-84184","title":"Indian blockchain and AI services firm: 10-year company archive","headline":"Indian blockchain and AI services firm: 10 years of email, client requirements, quotations, invoices and HR records","description":"Ten years of company records from an Indian enterprise blockchain and AI development firm whose work spans custom coins, wallets, tokens, exchanges, stablecoins and blockchain audits. The archive holds email, customer requirements, quotations, invoices, and HR and payroll records for staff, about 5 TB in file storage.","category":"Company operations archive","sector":"Software & internet","subsector":"Blockchain and AI development services","dataTypes":["Email","Documents","Tabular / structured"],"useCases":["Enterprise workflow agents","Pre-training","Domain fine-tuning"],"origin":"First-party company records","regions":["South Asia"],"languages":["English"],"sizeBand":"1-10 TB","count":"~5 TB","volume":"~5 TB","years":"2026","yearsOfHistory":"10","growing":"Yes","systems":["File storage","Email"],"contents":["Email","Customer requirements","Quotations","Invoices","HR and payroll records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"2-6 weeks","stage":"Summary available","scale":"Mid-size company","employees":"51–200","industry":"Blockchain and AI development services","tools":[["email","Email"],["files","File storage"]],"span":"10 years","volumeHead":"5 TB","sizeTB":5.0,"fromYear":null,"toYear":2026,"yearsNum":10.0,"tags":["company archive","blockchain","crypto","software services","email","invoices","quotations","india"],"shortTitle":"Blockchain and AI services firm"},{"ref":"MK-85040","title":"Market-research planning, execution and consumption workflows","headline":"End-to-end market-research workflows: client problem to research plan, survey design, execution and client consumption","description":"Records from a market-research platform of how expert researchers interpret a client's problem statement, turn it into a well-formed research plan, design instruments such as surveys, execute the research, and how the client then consumes the results. The seller describes it as a tool-agnostic, end-to-end dataset for the market-research vertical, useful for generating high-quality surveys and other research instruments and for modelling how firms use research outputs. Data comes from platform users in the US and APAC from June 2026 onward and accrues continuously; licensing rights are not yet in place (an opt-in model for users is under consideration). Labels are present but not described; a sample is offered and delivery can be by API or flat files.","category":"Expert workflows","sector":"Marketing, media & publishing","subsector":"Market research platform","dataTypes":["Documents","Text","Tabular / structured"],"useCases":["SFT / instruction tuning","Enterprise workflow agents","Domain fine-tuning","Evaluation / benchmarks"],"origin":"Platform / user-generated","regions":["North America"],"languages":[],"sizeBand":"","count":"","volume":"","years":"2026","yearsOfHistory":"","growing":"Yes","systems":["Research platform"],"contents":["Client problem statements","Research plans","Survey and research instrument designs","Research execution records","Client consumption of research outputs"],"formats":["Several (not specified)"],"delivery":"API; Flat files","labelled":"Yes","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"Available","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Market research platform","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":2026,"toYear":2026,"yearsNum":null,"tags":["market research","surveys","research plans","expert workflows","questionnaire design","post-training"],"shortTitle":"Market-research workflows"},{"ref":"MK-85378","title":"Multi-agent web and browser RL environment with verifiable tasks","headline":"RL environment for web and browser agents: aggregation, informational and booking tasks, scored by deterministic verifiers and LLM judges","description":"A multi-agent web and browser environment, written in TypeScript and Rust, in which AI builds custom interfaces for navigating the web. Task families cover aggregation (shopping and comparison), informational queries and agentic completion such as booking and confirmation. Agents act through browser and computer use and API calls. Rewards combine LLM-as-judge scoring with deterministic verifiers. Tasks are packaged in Harbor format and calibrated against leading models; a deterministic web-search simulator is in progress. Delivery covers source, binaries and tasks, and a demo is available on request. Task count has not been stated yet.","category":"RL environment / benchmark","sector":"AI & data","subsector":"Agent training environments","dataTypes":["Code","Tasks & environments"],"useCases":["Computer-use & tool-use agents","Reinforcement learning","Agent evaluation"],"origin":"Expert-created","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":["Aggregation tasks (shopping, comparison)","Informational tasks","Agentic completion tasks (booking, confirmation)","Deterministic verifiers","LLM-as-judge rewards"],"formats":["Harbor tasks","Source code","Binaries"],"delivery":"Source, binaries and tasks","labelled":"Yes","labelTypes":["Verifier and LLM-judge rewards"],"personalData":"","deidentification":"","rights":"Owner-held","sample":"On request","readiness":"1-2 weeks","stage":"Summary available","scale":"","employees":"","industry":"Agent training environments","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["rl environment","browser agents","computer use","web tasks","verifiers","llm judge"],"shortTitle":"Web-browsing agent RL environment"},{"ref":"MK-86179","title":"Early-stage online retailer: full operating records","headline":"Early-stage retail and e-commerce business: under 2 years of email, CRM, accounting, code, orders and contracts","description":"Records of a small retail and e-commerce business under 2 years old with under 100 customers: email, CRM, accounting, code, files, project boards, orders, a database and contracts. The seller states it owns the data. Volumes are not stated.","category":"Company operations archive","sector":"Retail & e-commerce","subsector":"Early-stage online retailer","dataTypes":["Email","Tabular / structured","Code","Documents"],"useCases":["Enterprise workflow agents","Computer-use & tool-use agents","Coding models"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"<2","growing":"","systems":["Email","CRM","Accounting","Code repository","File storage","Project management","Database"],"contents":["Email","CRM records","Accounting","Code","Files","Project boards","Orders","Contracts"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Early-stage online retailer","tools":[["tickets","Tickets"],["git","Git"],["email","Email"],["crm","CRM"],["database","Database"],["accounting","Accounting"],["files","File storage"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["e-commerce","retail","startup","orders","crm","accounting"],"shortTitle":"New online retailer ops archive"},{"ref":"MK-86302","title":"US product-design and engineering studio: native CAD archive with revisions","headline":"Thousands of product-design projects since 2014: native SolidWorks with feature trees, numbered revisions and linked client feedback","description":"Archive of a US product-design and engineering business covering thousands of projects since 2014 across automotive, medical, consumer, architecture, sports and industrial work. Contents include native SolidWorks files with editable feature trees and assemblies, STEP, AutoCAD and 2D drawings, 3D-print files, prototype photos and videos, project folders with numbered revisions, email (since 2022), and client PowerPoint feedback linked to CAD revisions. About 1 TB. Designs carry client IP, so licensable scope is still to be confirmed.","category":"Expert workflows","sector":"Industrial & engineering","subsector":"Product design and engineering services","dataTypes":["3D / CAD","Documents","Image","Video","Email"],"useCases":["Domain fine-tuning","Vision & perception","Enterprise workflow agents"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"1-10 TB","count":"Thousands of projects","volume":"~1 TB; thousands of projects; email since 2022","years":"2014","yearsOfHistory":"","growing":"","systems":["Email","File storage"],"contents":["Native SolidWorks parts and assemblies with feature trees","STEP and AutoCAD files","2D drawings","3D-print files","Prototype photos and videos","Numbered revision folders","Client PowerPoint feedback linked to revisions","Project email"],"formats":["SolidWorks","STEP","AutoCAD","3D-print files","Photos","Video","PowerPoint"],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Product design and engineering services","tools":[["autocad","AutoCAD"],["email","Email"],["files","File storage"]],"span":"","volumeHead":"Thousands of projects","sizeTB":1.0,"fromYear":2014,"toYear":null,"yearsNum":null,"tags":["cad","solidworks","product design","engineering","revision history","client feedback","3d"],"shortTitle":"Product-design CAD with revisions"},{"ref":"MK-87275","title":"Egocentric video of electricians and plumbers on Indian construction sites","headline":"Egocentric video of electricians and plumbers at work on Indian construction sites","description":"Head-camera video recorded on live construction sites in India, showing skilled trades at work: electricians assembling distribution boards and laying conduit wiring, and plumbers fixing taps, each performing tasks one after another. Partly labelled. Offered for training robot models on electrical and plumbing work; site conditions are dusty.","category":"Physical-world video","sector":"Construction & real estate","subsector":"Electrical and plumbing trades on construction sites","dataTypes":["Video"],"useCases":["Robotics / embodied AI","Vision & perception","World models"],"origin":"Purpose-collected (consented capture)","regions":["South Asia"],"languages":[],"sizeBand":"10-100 GB","count":"~10 GB video","volume":"~10 GB","years":"2026","yearsOfHistory":"","growing":"","systems":[],"contents":["Electrical board assembly","Conduit wiring installation","Tap fitting and repair","Sequential task footage"],"formats":["MP4"],"delivery":"Download","labelled":"Partly","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"Not available","readiness":"Within 2 weeks","stage":"Summary available","scale":"","employees":"","industry":"Electrical and plumbing trades on construction sites","tools":[],"span":"","volumeHead":"10 GB video","sizeTB":0.01,"fromYear":2026,"toYear":2026,"yearsNum":null,"tags":["egocentric video","construction","electricians","plumbers","skilled trades","india","robotics"],"shortTitle":"Electrician and plumber video"},{"ref":"MK-87404","title":"AI SaaS company: 2-4 year company archive","headline":"AI SaaS company with a large customer base: email, Slack, GitHub, call recordings, Drive files, contracts and HR over 2-4 years","description":"Company data from an AI software/SaaS business with a customer base in the 100K-999K range and 2-4 years of history. Systems listed: email, team chat, code repositories, call recordings, file storage, contracts and HR records. Volumes are not yet stated.","category":"Company operations archive","sector":"Software & SaaS","subsector":"AI SaaS","dataTypes":["Email","Chat & messaging","Code","Audio / speech","Documents"],"useCases":["Coding models","Enterprise workflow agents","Speech & audio models","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"2-4 years","yearsOfHistory":"2-4","growing":"","systems":["Email (Gmail)","Team chat (Slack)","GitHub","Call recordings","File storage (Google Drive)","Contracts","HR"],"contents":["Email","Slack chat","GitHub code","Call recordings","Drive files","Contracts","HR records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"AI SaaS","tools":[["slack","Slack"],["github","GitHub"],["gmail","Gmail"],["googledrive","Google Drive"],["calls","Calls"],["hr","HR & payroll"]],"span":"4 years","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["saas","ai","code","call recordings","slack"],"shortTitle":"AI SaaS startup company archive"},{"ref":"MK-87783","title":"Arabic-language consumer app and marketplace: 2-4 year company archive","headline":"Small consumer app and marketplace software company: email, team chat, CRM, accounting, support, code, recordings, contracts and HR, mainly in Arabic","description":"Operating data from a small software company running a consumer app and marketplace, covering 2-4 years. The seller lists Gmail, Slack, Salesforce, QuickBooks, customer support, source code, recordings, contracts and HR records. The primary language is Arabic. Volumes have not been provided yet.","category":"Company operations archive","sector":"Software & internet","subsector":"Consumer app and marketplace","dataTypes":["Email","Chat & messaging","Tabular / structured","Code","Audio / speech","Documents"],"useCases":["Enterprise workflow agents","Domain fine-tuning","Coding models","Arabic language models"],"origin":"First-party company records","regions":["Not stated"],"languages":["Arabic"],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"2-4","growing":"","systems":["Gmail","Slack","Salesforce","QuickBooks","Support desk","Code repository","Recordings","HR"],"contents":["Email","Team chat","CRM records","Accounting records","Support conversations","Source code","Recordings","Contracts","HR records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Timing to confirm","stage":"Summary available","scale":"Growing company","employees":"11–50","industry":"Consumer app and marketplace","tools":[["slack","Slack"],["salesforce","Salesforce"],["gmail","Gmail"],["quickbooks","QuickBooks"],["git","Git"],["hr","HR & payroll"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["arabic","consumer app","marketplace","company archive","slack","crm","code"],"shortTitle":"Arabic marketplace app archive"},{"ref":"MK-88117","title":"Pool service and repair operations data","headline":"US pool service and repair business: equipment, service history, water chemistry, repairs, quotes and outcomes","description":"Operating data of an owner-run US residential pool service and repair business: equipment records, service history, technician observations, water-chemistry readings, repairs, quoting, customer approvals and job outcomes. The data is increasingly being structured and the owner is open to guidance on what to collect going forward. The owner is also involved with other home-services businesses as an investor or consultant.","category":"Company operations archive","sector":"Consumer services","subsector":"Pool service and repair","dataTypes":["Tabular / structured","Text"],"useCases":["Domain fine-tuning","Enterprise workflow agents","Forecasting & time series"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"Yes","systems":["Field service management"],"contents":["Pool equipment records","Service history","Technician observations","Water-chemistry readings","Repair records","Quotes","Customer approvals","Job outcomes"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Pool service and repair","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["pool service","home services","field service","water chemistry","repairs","quotes","technician notes"],"shortTitle":"Pool service operations"},{"ref":"MK-88496","title":"Insurance pricing and risk data","headline":"Specialty insurance data already sold to carriers and brokers, from pricing to risk-modelling data","description":"Insurance data that the seller already sells to insurance carriers and retail brokers, ranging from pricing data to what the seller calls 'world-building' (risk-modelling) data. The seller is open to licensing to AI labs; an inventory has not yet been provided.","category":"Financial & transactional","sector":"Financial services & insurance","subsector":"Specialty insurance","dataTypes":["Tabular / structured"],"useCases":["Domain fine-tuning","Forecasting & time series","World models"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":["Insurance pricing data","Risk-modelling ('world-building') data","Data products sold to carriers and retail brokers"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Specialty insurance","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["insurance","pricing","risk modelling","carriers","brokers","specialty insurance"],"shortTitle":"Insurance pricing and risk"},{"ref":"MK-89696","title":"Vehicle-maintenance web app codebase","headline":"Codebase of a small vehicle-maintenance web app, under 2 years old","description":"Codebase of a small web application for vehicle maintenance, hosted on a cloud deployment platform and under 2 years old. The seller's details are inconsistent and volumes are not stated.","category":"Code & software","sector":"Software & internet","subsector":"Vehicle-maintenance web app","dataTypes":["Code"],"useCases":["Coding models"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"<2","growing":"","systems":["Hosting"],"contents":["Web app source code"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Vehicle-maintenance web app","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["codebase","web app","vehicle maintenance","automotive"],"shortTitle":"Vehicle-maintenance app codebase"},{"ref":"MK-90213","title":"AI video, image and games startup: 4-year archive","headline":"Small AI video/image, agents and games company: ~5 TB of email, community chat, code, boards, database and media assets","description":"Four years of operating data from a small company building AI video and image tools, agents, games and web3 products. About 5 TB across email, community chat, code repositories, file storage, project boards and a PostgreSQL database, plus images, videos, game assets, documents and code. Which code can be shared is still to be confirmed.","category":"Company operations archive","sector":"Software & internet","subsector":"AI video/image, agents and games","dataTypes":["Email","Chat & messaging","Code","Documents","Image","Video","Tabular / structured"],"useCases":["Enterprise workflow agents","Coding models","Vision & perception"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"1-10 TB","count":"~5 TB","volume":"~5 TB over 4 years","years":"","yearsOfHistory":"4","growing":"","systems":["Email (Gmail)","Team chat (Discord)","Git (GitHub)","File storage (Google Drive)","Project management (Trello)","Database (PostgreSQL)"],"contents":["Email","Discord chat","Code repositories","Drive documents","Trello boards","PostgreSQL database","Images and videos","Game assets"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"AI video/image, agents and games","tools":[["discord","Discord"],["github","GitHub"],["trello","Trello"],["gmail","Gmail"],["googledrive","Google Drive"],["database","Database"]],"span":"4 years","volumeHead":"5 TB","sizeTB":5.0,"fromYear":null,"toYear":null,"yearsNum":4.0,"tags":["ai video","image generation","games","game assets","web3","discord","github"],"shortTitle":"AI media and games archive"},{"ref":"MK-91055","title":"Haptic-glove teleoperation data","headline":"Robot teleoperation data from haptic gloves: force and tactile signals with hand motion","description":"Robot teleoperation data captured with haptic gloves, recording force and tactile signals alongside hand motion. Offered through the glove manufacturer's team, which can capture data; the volume already held, tasks and formats have not yet been stated.","category":"Robotics","sector":"AI & data","subsector":"Robotics teleoperation / haptics","dataTypes":["Sensor / telemetry"],"useCases":["Robotics / embodied AI","World models"],"origin":"Purpose-collected (consented capture)","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"","growing":"","systems":["Haptic gloves"],"contents":["Force data","Tactile data","Hand motion tracking","Robot teleoperation sessions"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"To be confirmed","sample":"","readiness":"Collected to order","stage":"Summary available","scale":"","employees":"","industry":"Robotics teleoperation / haptics","tools":[],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["robotics","teleoperation","haptics","tactile","force sensing","hand motion","manipulation"],"shortTitle":"Haptic-glove teleoperation"},{"ref":"MK-91257","title":"Writing-process data: how people draft and revise, plus final text","headline":"Writing-process metrics with the final text, captured across email, docs, notes and publishing tools from consenting writers (~4 GB)","description":"Records how people write: process metrics (drafting and revision behaviour) together with the final text, captured from email, documents, note-taking and publishing tools. Writers consent at sign-up; third-party text can be stripped.","category":"Human preference / evaluation","sector":"AI & data","subsector":"Writing behaviour","dataTypes":["Text"],"useCases":["Preference / RLHF","SFT / instruction tuning"],"origin":"Purpose-collected (consented capture)","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"~4 GB","years":"Feb 2026-present","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held","sample":"Available","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Writing behaviour","tools":[],"span":"","volumeHead":"4 GB","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["writing","revisions","human process","text"],"shortTitle":"Writing-process data"},{"ref":"MK-91321","title":"AI-security company (red-teaming, guardrails): 3 years of code, wikis, 300 mailboxes, Slack, CRM, proposals","headline":"US AI-security company working on red-teaming, guardrails and AI detection and response: all code, Notion wikis, 300 mailboxes, Slack, CRM, proposals and legal documents","description":"Operating record of a US AI-security company that red-teams models and builds guardrails and AI detection and response: full code history, internal wikis, email across 300 mailboxes, team chat, CRM, proposals and legal documents.","category":"Company operations archive","sector":"AI & data","subsector":"AI security","dataTypes":["Code","Email","Team chat","Documents","Tabular / structured"],"useCases":["Cybersecurity & red-teaming","Enterprise workflow agents","Coding models"],"origin":"First-party company records","regions":["North America"],"languages":[],"sizeBand":"","count":"","volume":"300 mailboxes; volume stated as ~100 TB (to verify)","years":"~3 years","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Medium","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"Export 1-2 weeks","stage":"Summary available","scale":"Mid-size company","employees":"51–200","industry":"AI security","tools":[["slack","Slack"],["notion","Notion"]],"span":"3 years","volumeHead":"300 mailboxes","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["ai security","red teaming","guardrails","code","email"],"shortTitle":"AI-security company archive"},{"ref":"MK-94038","title":"Egocentric smart-glasses video from car and tyre workshops (~2,000 hours)","headline":"~2,000 hours (~16 TB) of first-person smart-glasses video of real work in owned car and tyre workshops","description":"Egocentric video captured with smart glasses in car and tyre workshops the owner runs, for physical-AI and dexterity models. Unlabelled. Delivery via cloud bucket or API.","category":"Physical-world video","sector":"Automotive & transport","subsector":"Car and tyre workshops","dataTypes":["Video"],"useCases":["Robotics / embodied AI","World models"],"origin":"Purpose-collected (consented capture)","regions":["Southeast Asia"],"languages":[],"sizeBand":"10-100 TB","count":"~2,000 hours","volume":"~2,000 hours; ~16 TB","years":"","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":["MP4"],"delivery":"","labelled":"No","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Car and tyre workshops","tools":[],"span":"","volumeHead":"2K hours","sizeTB":16.0,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["egocentric","video","workshop","dexterity"],"shortTitle":"Tyre workshop smart-glasses video"},{"ref":"MK-94119","title":"Consumer AI assistant usage traces","headline":"Usage traces from a consumer AI assistant product","description":"Interaction traces generated by end users of the seller's own consumer AI product. The data is first-party platform data; trace counts, formats and time span have been requested but not yet provided.","category":"Agent traces","sector":"AI & data","subsector":"Consumer AI application","dataTypes":["Chat & messaging","Agent trajectories"],"useCases":["SFT / instruction tuning","Preference / RLHF","Evaluation / benchmarks","Computer-use & tool-use agents"],"origin":"Platform / user-generated","regions":["Not stated"],"languages":[],"sizeBand":"","count":"~300-1,000 requests","volume":"~300-1,000 requests","years":"","yearsOfHistory":"","growing":"","systems":["Application database"],"contents":["Consumer AI product usage traces","Consumer questions to the assistant","Consumer task requests","Product interaction logs","Session-level usage records","User interactions with an AI assistant"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"Moderate","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Consumer AI application","tools":[["database","Database"]],"span":"","volumeHead":"~300-1,000 requests","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["consumer ai","usage traces","assistant","chat logs","user interactions","first-party"],"shortTitle":"Consumer AI assistant traces"},{"ref":"MK-96485","title":"Human aesthetic ratings on design references (~23.6K ratings)","headline":"~23.6K human aesthetic ratings and ~500 annotations on ~2.5K design references (~0.9 GB)","description":"Human aesthetic judgements on design references: ~2.5K references, ~23.6K ratings and ~500 annotations, collected since March 2026.","category":"Human preference / evaluation","sector":"AI & data","subsector":"Design preference","dataTypes":["Image","Tabular / structured"],"useCases":["Preference / RLHF","Evaluation / benchmarks"],"origin":"Platform / user-generated","regions":["Not stated"],"languages":[],"sizeBand":"","count":"~23.6K ratings","volume":"~23.6K ratings; ~2.5K references; ~0.9 GB","years":"Mar 2026-present","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"Yes","labelTypes":[],"personalData":"Low","deidentification":"","rights":"To be confirmed","sample":"","readiness":"On request","stage":"Summary available","scale":"","employees":"","industry":"Design preference","tools":[],"span":"","volumeHead":"24K ratings","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["aesthetics","preference","design","ratings"],"shortTitle":"Design aesthetic ratings"},{"ref":"MK-96629","title":"US political-news newsletter: ~10-20M subscriber engagement histories (opens, clicks, interests) over 3 years","headline":"Email engagement history for a large US political-news newsletter audience: emails received, opens, clicks and interest signals, offered anonymised","description":"Engagement records from a US political and news email newsletter publisher: per-subscriber history of emails received, opens, clicks and interest signals over about three years. Offered only in anonymised form; consent basis under review.","category":"Consumer behaviour","sector":"Media & entertainment","subsector":"Newsletters","dataTypes":["Tabular / structured"],"useCases":["Recommendation","Consumer behaviour models"],"origin":"Platform / user-generated","regions":["North America"],"languages":[],"sizeBand":"","count":"","volume":"~10-20M subscriber records with engagement history","years":"~3 years","yearsOfHistory":"","growing":"","systems":[],"contents":[],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"Required before release","rights":"Owner-held, conditions to confirm","sample":"","readiness":"On request","stage":"Summary available","scale":"Small team","employees":"1–10","industry":"Newsletters","tools":[],"span":"3 years","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["newsletter","engagement","email","media"],"shortTitle":"Political newsletter engagement"},{"ref":"MK-96684","title":"Wholesale distributor: 10-19 years of email, chat, CRM, ERP, orders, code and recordings","headline":"Wholesale and distribution company with a large customer base: email, team chat, CRM, ERP, support tickets, code, meeting recordings, orders and field operations","description":"Operating data from a wholesale and distribution company with 10-19 years of history and a customer base in the hundreds of thousands. Systems listed are Gmail, Slack, CRM, NetSuite ERP, support tickets, GitHub, Google Meet recordings, file storage, an internal wiki, orders and point of sale, field operations and a database. Volumes have not been provided yet.","category":"Company operations archive","sector":"Logistics, distribution & wholesale","subsector":"Wholesale distribution","dataTypes":["Email","Chat & messaging","Tabular / structured","Code","Video","Audio / speech","Documents"],"useCases":["Enterprise workflow agents","Forecasting & time series","Coding models","Domain fine-tuning"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"","count":"","volume":"","years":"","yearsOfHistory":"10-19","growing":"","systems":["Gmail","Slack","CRM","ERP (NetSuite)","Support tickets","GitHub","Meeting recordings (Google Meet)","File storage","Wiki","Orders / POS","Field operations","Database"],"contents":["Email","Team chat","CRM records","ERP records","Support tickets","Source code","Meeting recordings","Wiki pages","Orders and POS records","Field operations records"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"","deidentification":"","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"Mid-size company","employees":"51–200","industry":"Wholesale distribution","tools":[["slack","Slack"],["github","GitHub"],["gmail","Gmail"],["erp","NetSuite"],["tickets","Tickets"],["crm","CRM"],["wiki","Docs & wiki"],["database","Database"],["files","File storage"]],"span":"","volumeHead":"","sizeTB":null,"fromYear":null,"toYear":null,"yearsNum":null,"tags":["wholesale","distribution","erp","crm","orders","support tickets","meeting recordings","code"],"shortTitle":"Wholesale distributor ops archive"},{"ref":"MK-97275","title":"Merchant-built fashion lookbooks: 700K+ outfit bundles","headline":"700K+ merchant-created fashion lookbooks showing which products go together, with product information, 2017-2025","description":"Lookbook bundles created by merchants in a shop-the-look app for a major e-commerce platform. Each of the 700K+ lookbooks groups fashion items that go together (complementary items) and includes product information. Pitched for fashion recommendation and styling AI. Covers 2017-2025 and is delivered as zipped CSV. Merchant entries were made under the app's terms of service up to 2025; the seller notes the platform's terms later changed to prohibit using merchant data for AI training, so the rights scope needs confirming. Stated not to contain sensitive data. A sample is available.","category":"Consumer behaviour","sector":"Retail & e-commerce","subsector":"Fashion lookbook app for online merchants","dataTypes":["Tabular / structured","Text"],"useCases":["Recommendation and styling models","Domain fine-tuning"],"origin":"Platform / user-generated","regions":["Not stated"],"languages":[],"sizeBand":"","count":"700K+ lookbooks","volume":"700K+ lookbook bundles with product information","years":"2017–2025","yearsOfHistory":"8","growing":"","systems":["E-commerce platform app"],"contents":["Merchant-created lookbook bundles","Complementary fashion item groupings","Product information"],"formats":["CSV (zip)"],"delivery":"","labelled":"Yes","labelTypes":[],"personalData":"Low","deidentification":"","rights":"Owner-held, conditions to confirm","sample":"Available","readiness":"1-2 weeks","stage":"Summary available","scale":"","employees":"","industry":"Fashion lookbook app","tools":[["shop","Commerce"]],"span":"8 years","volumeHead":"700K lookbooks","sizeTB":null,"fromYear":2017,"toYear":2025,"yearsNum":8.0,"tags":["fashion","lookbooks","outfit recommendation","complementary items","styling","e-commerce","merchant content"],"shortTitle":"Merchant fashion lookbook bundles"},{"ref":"MK-97903","title":"Venture-backed software startup: email, code, tickets and product databases","headline":"Software startup company record: ~3.7 TB email and docs, ~6M lines of code, ~14K tickets, ~300M database records","description":"The full working record of a venture-backed software startup: staff email and document storage, about three years of team chat, ~65 private code repositories, issue-tracker tickets, design files, HR records and product databases with telemetry and in-product chat messages. Counts are owner-reported; a detailed inventory is pending.","category":"Code & software","sector":"Software & internet","subsector":"Venture-backed software startup","dataTypes":["Email","Documents","Chat & messaging","Code","Tabular / structured","Sensor / telemetry"],"useCases":["Enterprise workflow agents","Coding models","Pre-training","Retrieval / RAG"],"origin":"First-party company records","regions":["Not stated"],"languages":[],"sizeBand":"1-10 TB","count":"~300M database records","volume":"~3.7 TB email and documents; ~6M lines of code in ~65 repos; ~14K tickets; ~300M database records","years":"~3 years","yearsOfHistory":"3","growing":"","systems":["Email","File storage","Team chat","Git","Issue tracker","Design files","HR system","Database"],"contents":["Staff email","Shared documents","Team chat","Private code repositories","Issue-tracker tickets","Design frames","HR records","Product telemetry","In-product chat messages"],"formats":[],"delivery":"","labelled":"","labelTypes":[],"personalData":"High","deidentification":"Planned / on request","rights":"Owner-held","sample":"","readiness":"On request","stage":"Summary available","scale":"Large organisation","employees":"200+","industry":"Venture-backed software startup","tools":[["chat","Team chat"],["tickets","Tickets"],["git","Git"],["email","Email"],["database","Database"],["design","Design files"],["hr","HR & payroll"],["files","File storage"]],"span":"3 years","volumeHead":"300M database records","sizeTB":3.7,"fromYear":null,"toYear":null,"yearsNum":3.0,"tags":["startup","codebase","email","slack","tickets","telemetry","company archive"],"shortTitle":"Venture-backed startup archive"}]}