To transcribe an audio recording, save it as a file and send it to the Potok bot on Telegram, which accepts audio and video up to 2 GB. One hour of mp3 at 128 kbit/s weighs about 57.6 MB, so even long recordings go through whole, with no need to cut them into parts.
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
- One hour of mp3 at 128 kbit/s is about 57.6 MB: 128,000 bits per second times 3,600 seconds, divided by 8.
- Uncompressed stereo WAV at 44.1 kHz and 16 bit is about 635 MB an hour, so roughly three hours fit under 2 GB.
- The bot reads the format itself; mp3, m4a, wav, ogg, opus, flac, aac, amr and wma are accepted.
- Switch on speaker labels with more than one voice; labels are numbers such as Speaker 1, not names.
How to do it
- Save the recording as a file. Take it from the recorder app, the call history, a messenger or the page of a meeting recording. A link with closed access has to be downloaded first.
- Listen to the first seconds. If speech is audible, recognition will cope. Voices drowned in noise leave gaps in the text, and that is a property of the recording.
- Send the file to the bot. Open @potok_now_bot in Telegram and attach the file. For very long recordings use the Upload button of the mini app.
- Turn on speaker labels. With two or more voices, labels give each reply a Speaker number and a timestamp.
- Check names and numbers. Find the place by its timestamp and listen again before a name or an amount goes into a document.
What the result looks like
[00:00] Speaker 1: I am noting the results: the delivery moves to Thursday.
[00:11] Speaker 2: Did the warehouse confirm the time? I need it in writing.
[00:27] Speaker 1: Confirmed. I will send a message before the end of the day.
[00:41] Speaker 2: Then we fix it: Thursday, second half of the day.
Sample text written by us to show the layout of a transcript. It is not the output of the recognition engine and not a real recording.
| 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 link with closed access | The bot cannot open someone else's storage, so no text comes back. | Download the recording and send the file itself. |
| Cutting a long file into pieces without need | Speaker numbering breaks and context is lost. | Send it whole; up to 2 GB passes as it is. |
| Taking Speaker 1 and Speaker 2 for names | Replies get attributed to the wrong people. | Write the real names in by hand after the transcript. |
| Not checking figures against the sound | A wrong amount or surname travels on into a document. | Replay the spot by its timestamp. |
More
See the full list of languages, the tools and the glossary.
Frequently asked questions
Which audio formats does the bot accept?
It accepts mp3, m4a, wav, ogg, opus, flac, aac, amr and wma, plus video files with a sound track. You do not need to convert anything beforehand.
How much does an hour of audio weigh?
An hour of mp3 at 128 kbit/s is about 57.6 MB, and an hour of uncompressed stereo WAV is about 635 MB. Both stay below the 2 GB limit.
Can the text separate speakers?
Yes, switch on speaker labels. Replies are marked Speaker 1, Speaker 2 and so on, and the bot does not guess names.
What if the recording is quiet or noisy?
Send it as it is and check the text against the sound by timestamps. Next time, place the recorder closer to the speaker.
Sources
Each source was opened by hand on the date shown. Pages change, so open the source if a detail matters to you.
- MPEG-1 Audio Layer III / MP3 (Library of Congress), Library of Congress · checked
- WAVE audio file format (Library of Congress), Library of Congress · 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