French speech to text in Potok is a recommended language. Published large-v3 tests give word error rates between 5.3 and 6.31, from several sources, not our own measurement. French spelling hides many sounds, so errors tend to be about spelling rather than hearing.
Transcribe a recording
Send your voice message, audio or video to @potok_now_bot on Telegram and the text comes back in the chat.
Open @potok_now_bot in Telegram
Files up to 2 GB · export to TXT, MD, DOCX, PDF, SRT and VTT · speaker labels and timestamps · about 30 languages
Key facts
- Several published sources agree within about one point, which is a steadier picture than a single paper.
- Many French words sound alike but are spelled differently, and context decides which one the text uses.
- Agreement endings such as plural s or e often cannot be heard, so they depend on grammar.
- Canadian and European speakers differ in pronunciation, and the figures do not separate them.
- Liaison links the end of one word to the start of the next and can blur word boundaries in fast speech.
What a transcript looks like
[00:00] Speaker 1: Commençons la réunion. D'abord, nous passerons en revue les résultats de la semaine dernière, puis nous planifierons les prochaines étapes.Meaning in English: Let's start the meeting. First, we'll review last week's results, then plan the next steps.
Written by us to show the script and the layout of a transcript. It is not the output of the recognition engine.
Published quality
| Language | Verdict | Published result | Notes |
|---|---|---|---|
| French (Français) | Recommended | word error rate 5.3% (large-v3, FLEURS) Canary-1B-v2 paper, Whale paper, TheWhisper benchmark | Published tests show a low error rate for this language. Published large-v3 results range from 5.3 to 6.31. |
These numbers come from published tests by other people, not from tests by Potok. Your own audio, with its noise, accents and microphone, will differ.
How to do it
- Choose your speaker. Note whether the voice is from France, Canada or elsewhere.
- Send the clip. Open @potok_now_bot and attach it.
- Check agreement. Look at plurals, past participles and feminine forms.
- Send the rest. Up to 2 GB; use Upload for large files.
- Export. Pick TXT, DOCX, PDF, MD, SRT or VTT.
- Keep the source audio. Hold on to the original so a doubtful liaison or ending can be replayed later.
| Feature | Details |
|---|---|
| Where it works | Telegram bot @potok_now_bot; large files go through the bot's mini app |
| File size | Up to 2 GB |
| Export | TXT, MD, DOCX, PDF, SRT, VTT |
| Speakers and time | Speaker labels and timestamps can be switched on |
| Languages | About 30 languages with good published results; a few more with warnings |
| Translation | The transcript is not translated into another language |
| Dictation | A separate Potok desktop app lets you dictate text into any program; it recognises Russian speech |
Common mistakes
| Mistake | What happens | Fix |
|---|---|---|
| Treating spelling slips as hearing failures | The sound was right but the spelling differs. | Proofread by grammar. |
| Ignoring silent endings | Plural markers get lost. | Check them against the sentence. |
| Trusting 5.3 for a phone line | Phone audio is harder. | Test a phone clip. |
| Reading the range as our score | All figures were published by others. | Cite them as such. |
| Expecting identical numbers from every source | The tests differ in audio and text rules. | Quote the range, not one value. |
More
See the full list of languages, the tools and the glossary.
Frequently asked questions
What is the published range?
From 5.3 to 6.31 word error rate for large-v3, taken from several published tests.
Why are there several numbers?
Different tests use different audio and text rules, so they do not match exactly.
Does it support Quebec French?
The tests do not separate it from other speech, so try a short sample from the speaker.
Is the transcript in French?
Yes, it stays in French because the bot does not translate.
How big can the file be?
Files up to 2 GB are accepted, with the Upload button for the largest recordings.
Sources
Each source was opened by hand on the date shown. Pages change, so open the source if a detail matters to you.
- Word error rate (Wikipedia), Wikipedia · checked
- FLEURS speech dataset (Google, Hugging Face), Hugging Face · checked
- Canary-1B-v2 paper, Table 11 (Whisper large-v3 on FLEURS) · checked
- Whale paper, Table 5 (Whisper large-v3 on FLEURS) · checked
- TheWhisper benchmark (turbo, average over FLEURS, MLS, Common Voice) · checked
Try it on your own recording
Send your voice message, audio or video to @potok_now_bot on Telegram and the text comes back in the chat.
Open @potok_now_bot in Telegram