Potok converts a Bulgarian audio file to text when you send it to the Telegram bot. Published tests of others on the large-v3 model, not our measurements, give a word error rate of 12.86 for Bulgarian, which places it among the languages with good results. Files can be up to 2 GB.
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
- Word error rate counts the share of words that differ from a correct reference text, so lower is better.
- Bulgarian is written in the Cyrillic script, so the export should open in an editor with a Unicode font.
- The score is an average over test speech; noisy phone recordings usually score worse.
- Speaker labels are numbers such as Speaker 1, never real names.
What a transcript looks like
[00:00] Speaker 1: Да започнем срещата. Първо ще прегледаме резултатите от миналата седмица, а след това ще планираме следващите стъпки.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 |
|---|---|---|---|
| Bulgarian (Български) | Recommended | word error rate 12.86% (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 the best file you have. Use the original recording if it exists. A voice message passed through several apps has been compressed more than once and loses quiet detail.
- Start the bot. Open @potok_now_bot and send the file, or forward a voice message straight into the chat.
- Use Upload for large files. Up to 2 GB is allowed. The Upload button in the bot's mini app handles recordings too big for a normal chat attachment.
- Turn on speaker labels if needed. For an interview with two voices the numbered labels save you from guessing who spoke each line.
- Open the export in a Unicode editor. Choose DOCX or TXT and check that the letters display correctly before you share the text.
| 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 |
|---|---|---|
| Sending a heavily compressed forwarded clip | Quiet words vanish and the error count climbs. | Send the original file when possible. |
| Mixing two languages in one file | The part in the other language will read poorly. | Split the file where the language changes. |
| Treating 12.86 as a promise | The figure is an average from other people's tests. | Proofread names and numbers anyway. |
| Opening the text in an old editor | Cyrillic letters turn into question marks. | Use a current word processor, or save as DOCX. |
More
See the full list of languages, the tools and the glossary.
Frequently asked questions
Is Bulgarian a well supported language?
Its published word error rate of 12.86 on large-v3 puts it in the good group, although names and numbers still need a read-through.
Can I send a Bulgarian video?
Yes. Video files and YouTube links work like audio, within the 2 GB limit.
Does the text keep Bulgarian letters?
Yes, the transcript is written in Bulgarian. Open the file in an editor that supports Unicode so the letters show correctly.
Where does the error rate come from?
It comes from published tests by other people on the large-v3 model. We did not measure it on your audio.
How do I find who said what?
Switch on speaker labels. They appear as numbers, because the bot does not know 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