Mandarin Chinese is a good fit for Potok: send audio or video to the Telegram bot and receive Chinese text. Published tests of the large-v3 model give a character error rate of 7.8, down from 14.7 for the earlier large-v2 model. 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
- Chinese is written without spaces, so quality is a character error rate: 7.8 for large-v3 in published tests.
- The earlier large-v2 model scored 14.7 on the same kind of test, so the newer model is a clear step up.
- These are other groups' results, not measurements we made on your recordings.
- Check whether the output is in the character set you need, since readers may expect either simplified or traditional forms.
What a transcript looks like
[00:00] Speaker 1: 我们开始开会吧。首先回顾上周的结果，然后规划下一步。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 |
|---|---|---|---|
| Chinese (Mandarin) (中文（普通话）) | Recommended | character error rate 7.8% (large-v3, FLEURS) Whale paper, Whisper paper | Published tests show a low error rate for this language. Chinese is written without spaces, so quality is reported as a character error rate. The earlier large-v2 model scored 14.7. |
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
- Know your Mandarin. Standard broadcast speech is the easy case. Strong regional accents or a mix with Cantonese deserve a short test first.
- Send the file. Message @potok_now_bot with a voice note, audio, video or YouTube link. For recordings near 2 GB, use the Upload button.
- Switch on timestamps. Lines without spaces are harder to scan, so timestamps help you find a spot in a long text.
- Proof names and numbers. Personal names, company names and figures are where homophones cause mistakes.
- Export. SRT or VTT suit video subtitles; DOCX, PDF, TXT or MD suit reports.
| 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 |
|---|---|---|
| Comparing the 7.8 character rate with English word error rates | The units differ, so the comparison says little. | Compare character rates with character rates only. |
| Not checking homophones in names | A name can be written with the wrong characters while sounding right. | Confirm names against a source. |
| Expecting spaces between words | Chinese text has none, so word search tools behave differently. | Use character-based search in your editor. |
| Counting on a translation | The text stays Chinese and the bot does not translate. | Translate separately if needed. |
More
See the full list of languages, the tools and the glossary.
Frequently asked questions
How good is Mandarin transcription?
Published tests of large-v3 report a character error rate of 7.8. These are other groups' results, and your recording may behave differently.
Why a character rate and not a word rate?
Chinese has no spaces between words, so counting words needs extra rules. Characters give a simpler and fairer measure.
Will I get simplified or traditional characters?
The published tests do not say, so check a short sample and convert the text in your editor if you need the other form.
Does the bot handle Cantonese too?
Cantonese is a separate entry on our list with its own caution. This page covers Mandarin.
Can I export subtitles?
Yes. SRT and VTT files carry a time for every line, and you should proofread them before publishing.
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
- Whisper paper, Table 13 (large-v2 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