To convert Cantonese audio to text, send the file to the Potok bot on Telegram, knowing this language is borderline. Published tests of others on large-v3, not our measurements, give a character error rate of 9.8. Quality varies by model variant. Files up to 2 GB are accepted.
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
- Borderline: a vendor test reported much worse results for the faster turbo variant.
- The 9.8 is a character error rate from published tests of others, because Cantonese is written without spaces.
- Judge the result by wrong characters, not by wrong words.
- A short clip is the quickest way to see whether your speaker works well.
What a transcript looks like
[00:00] Speaker 1: 我哋開始開會啦。首先睇下上個禮拜嘅結果，跟住傾下一步點做。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 |
|---|---|---|---|
| Cantonese (粵語) | Recommended, borderline | character error rate 9.8% (large-v3, FLEURS) Whale paper | Published tests show an error rate close to our 15% line. Check a short sample first. Cantonese is written without spaces, so quality is reported as a character error rate. A vendor test reported much worse results for the faster turbo variant, so treat this language as borderline. |
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
- Run a short sample first. Because this language is borderline, cut thirty seconds of your recording and send it before the whole file. It costs little and shows what to expect.
- Open the bot. Open @potok_now_bot in Telegram and choose the chat.
- Send the full file. If the sample reads well, attach the whole recording, up to 2 GB. For a big file use the Upload button of the mini app.
- Read in characters, not words. Without spaces there are no word breaks, so scan line by line and mark every wrong character against the audio.
- Export. Save TXT or DOCX for editing, or SRT and VTT if you need subtitles with timestamps.
| 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 |
|---|---|---|
| Judging the result by word count | Text without spaces has no word boundaries to count. | Count characters, which is the metric the published figure uses. |
| Skipping the sample | A borderline language can fail on one voice and work on another. | Test thirty seconds before a long upload. |
| Reusing a figure from another model | Results for the faster turbo variant were reported as much worse. | Quote the model named in the source. |
| Sending a talk that switches into Mandarin | The recogniser follows one language per file. | Split the file where the language changes. |
More
See the full list of languages, the tools and the glossary.
Frequently asked questions
Why is Cantonese called borderline?
The published character error rate of 9.8 looks good, but one vendor test reported much worse results for the turbo variant, so the language needs care.
What does character error rate mean?
It counts wrong characters against the correct text. It suits languages written without spaces.
Can I send a long recording?
Yes, up to 2 GB. The Upload button of the bot's mini app is meant for big files.
Is the result checked by a person?
No. It is machine output, so read it against the recording before using it.
Can I get subtitles?
Yes. Export SRT or VTT, with timestamps switched on.
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
- Whisper speech recognition system (Wikipedia), Wikipedia · checked
- Whale paper, Table 5 (Whisper large-v3 on FLEURS) · 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