Potok converts Polish audio to text when you send a recording to the Telegram bot. Published tests of others on the large-v3 model, not our measurements, give a word error rate of 4.74 for Polish. Files up to 2 GB are accepted, and the text comes back in the chat or as a file.
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
- Polish has nine letters with marks, such as the barred l and the dotted z, and each must show correctly.
- Nouns and adjectives change their endings by case, so a wrong ending can still look like a real word.
- Long consonant clusters are common, and rapid speech is where they get blurred.
- A score of 4.74 comes from clean test speech, so a phone call will usually fare worse.
- Numbers are spoken as words with case endings, and the text may write them out or as digits depending on context.
What a transcript looks like
[00:00] Speaker 1: Zacznijmy spotkanie. Najpierw omówimy wyniki z zeszłego tygodnia, a potem zaplanujemy kolejne kroki.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 |
|---|---|---|---|
| Polish (Polski) | Recommended | word error rate 4.74% (large-v3, FLEURS) Canary-1B-v2 paper | Published tests show a low error rate for this language. |
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
- Pick a clean source. Record near the speaker if you can. Polish consonant clusters are hard to catch over noise.
- Open the bot. Open @potok_now_bot and send the file as audio or as a document.
- Use Upload for the biggest files. The mini app accepts up to 2 GB when a normal chat attachment will not.
- Check the marked letters. Look for the barred l, the nasal letters and the dotted z, and check that none was replaced with a plain letter.
- Check endings. Read a few sentences aloud against the audio. A changed ending changes the grammar.
- Export. Save as DOCX for sharing or SRT for captions.
| 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 |
|---|---|---|
| Opening the file in a tool without Polish fonts | Marked letters become boxes. | Use a current editor with Unicode support. |
| Skipping the ending check | A wrong case still reads smoothly. | Compare a sample with the audio. |
| Trusting surnames | Polish surnames have many spellings. | Verify each against a document. |
| Sending a call recording at low volume | Quiet speech loses consonants. | Raise the level before sending. |
More
See the full list of languages, the tools and the glossary.
Frequently asked questions
How good is Polish here?
The published word error rate is 4.74 on large-v3, which is a strong result, but proofreading still matters.
Do Polish letters survive the export?
Yes. The transcript is written in Polish, so open it in a program that supports Unicode.
Can I send a podcast episode?
Yes. Long files are fine up to 2 GB, with Upload for the biggest.
Did you measure the 4.74?
No. It comes from published tests by other people.
Can I label speakers?
Yes, as numbers such as Speaker 1, never as real names.
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
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