Datasets:
French Speech Dataset
1,131 hours of unscripted French recorded between 24 January and 26 February 2026, mostly by speakers in Algeria, Madagascar, Morocco, Rwanda and Benin: 1,012 hours of two-person conversation with each speaker on a separate channel, and 118 hours of single-speaker recordings. Machine transcripts with word-level timestamps cover 1,129 of the hours.
In stock. Download access within 24 hours of a signed order. The figures on this card were counted in our recording database on 18 November 2026.
This repository is the dataset card only. The audio is licensed commercially and isn't hosted on Hugging Face. Excerpts with their uncorrected machine transcripts play on the French dataset page, where you can also ask for the full sample recordings. To evaluate a purchase, request a written quote so the team can confirm available recordings, permitted use and total price for your brief.
Specifications
| Language | French |
| Speech | 1,012 hours of two-person conversation and 118 hours of single-speaker recordings |
| Volume | 1,131 hours in 2,088 recordings |
| Speakers | 82 registered speakers, 27 of whom recorded monologues. Their conversation partners did not register, so the dataset holds more than 82 voices; how many more has not been counted. |
| Hours by country | Algeria 487 h, Madagascar 202 h, Morocco 139 h, Rwanda 105 h, Benin 95 h, France 27 h, Côte d'Ivoire 25 h, Senegal 13 h, Burundi 12 h, Other or not stated 26 h |
| Recorded | 24 January 2026 to 26 February 2026 |
| Audio | WAV at 48 kHz, 16 or 24 bits depending on the session. Each conversation comes as a stereo file with one speaker per channel plus a separate file per speaker; single-speaker recordings are mono. |
| Devices | About 96% recorded on a laptop or desktop computer and about 4% on a phone, indoors in quiet rooms. |
| Transcripts | Machine transcripts for 1,129 of the 1,131 hours, not reviewed by a person: speaker labels, segment start and end times and word-level timestamps, as TXT and JSON. |
| Topics | More than 1,700 distinct topic labels, from business meetings, interviews and sales to football, music, religion and daily life. |
| Delivery | Download access within 24 hours of a signed order |
| License | Commercial use under a written license; voice cloning is excluded |
| Price | $35 per hour of audio when you license the full dataset; part of it is quoted on request |
What is in it
| Recording type | Recordings | Hours | Average length |
|---|---|---|---|
| Conversation, business role-play: Two speakers act out work situations: client meetings, deliveries, sales and interviews | 919 | 505.9 | 33 min |
| Conversation, open topics: Two speakers talk about daily life, football, music, education, religion and travel | 936 | 506.4 | 32 min |
| Monologue: One speaker talks about a topic they chose, from technology to history | 221 | 109.4 | 30 min |
| Single speaker, type not labeled: Mono recordings uploaded without a recording type | 12 | 8.9 | 44 min |
If you need speakers from France, Belgium, Switzerland or Canada, we can record them to order.
What makes French hard for speech recognition
French sounds different in France, the Maghreb, West Africa and Madagascar, and many speakers switch between French and Arabic, Malagasy or another local language. Spoken French also drops much of what its spelling keeps: the "ne" in negation usually disappears and "il y a" becomes "y'a", so one sound can map to several spellings.
Getting the data
Listen to the excerpts on the French dataset page and read the machine transcript lines under each one. Need another accent, more speakers or your own recording setup? SpeechData.ai also collects speech data to spec, or you can book a call.
Need something else?
- Call center speech data for call-center style recordings.
- Audio annotation services to transcribe or annotate audio you already have.
- Speech data pricing with published per-hour prices.
Related datasets
Compare all 9 African datasets by hours, speakers and price.
About SpeechData.ai
SpeechData.ai licenses conversational speech datasets in 59 languages and dialects and runs custom speech data collection. It's part of Spirelight. Browse the full catalog.
- Downloads last month
- 37