Indic two-speaker conversational speech

Indic two-speaker conversational speech: ~30K hours in 27 languages, 48 kHz speaker-isolated tracks, consented

CategorySpeech & conversation
SectorAI & data
IndustryConversational speech data collection
History2026
Volume30K hours (~30K hours)
Data typesAudio / speech
Source systemsFile storage
ReadinessReady now
LabelledPartly
SampleSample available to qualified buyers
Indicative price$900K for one non-exclusive licence, indicative and subject to the owner's sign-off
ReferenceMK-78006

Details are as stated by the owner. Who the owner is, the full description and the sample are shared with qualified buyers under our mutual NDA.

How to license this dataset

  1. Join Grepped Buyside with your work email. Our team reviews each buyer and lets them in.
  2. Sign our mutual NDA in the app. You then see the full listing and can open or request the sample.
  3. Ask for a quote or make an offer. Our team agrees scope and terms with the owner and with you, and the data is delivered once the licence is signed.

Grepped is a marketplace for data from real companies. Owners list their archives, codebases and records; AI labs and builders license them for post-training, evals and RL environments. More for buyers.

More speech & conversation data

All 11 speech & conversation datasets · Every category

Want this kind of data?

Qualified buyers see full listings, samples and quotes in Grepped Buyside. Or email e@grepped.ai.

Request access