Italian speech to text in Potok is a recommended language, and published tests of the large-v3 model report a word error rate of 2.31, one of the lowest on our list. That number comes from other people's work, not our measurements. Send voice notes, audio or video to the Telegram bot.
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
- A word error rate of 2.31 is close to what two careful humans might differ by, so most lines need only a light check.
- The published tests use standard Italian; strong regional speech, such as a dialect spoken at home, was not reported separately.
- Elided forms like l'amico or un'ora are single spoken units, so check that the apostrophes land in the right place.
- Double consonants change meaning in Italian, so a pair like pala and palla is worth a look in the output.
What a transcript looks like
[00:00] Speaker 1: Iniziamo la riunione. Prima esamineremo i risultati della settimana scorsa, poi pianificheremo i prossimi passi.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 |
|---|---|---|---|
| Italian (Italiano) | Recommended | word error rate 2.31% (large-v3, FLEURS) Canary-1B-v2 paper, TheWhisper benchmark | Published tests show a low error rate for this language. One of the lowest published error rates among the languages listed here. |
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
- Choose a clip with one main voice. An interview, a lecture or a voice message from a friend is a good test. Cut long lead-in silence if it is easy.
- Share it with the bot. Go to @potok_now_bot, then forward the message, attach the file, or paste a YouTube link.
- Use the mini app for big files. The bot takes files up to 2 GB. For the largest ones, press the Upload button in its mini app.
- Read the output against the sound. Short results appear in the chat; long ones arrive as a file. Listen to a minute and compare.
- Fix names and doubles. Surnames, towns and double-consonant words are the usual weak spots.
- Pick the export. TXT or DOCX for notes, SRT or 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 |
|---|---|---|
| Reading 2.31 as a promise for your file | It is a figure from published tests of others, on clean read speech. | Check one page of your own output before you trust the rest. |
| Sending a recording from a loud restaurant | Overlapping voices and clatter raise the number of wrong words. | Move the phone closer or record in a quieter spot next time. |
| Expecting dialect to match the standard | Speech with local vocabulary may be written in the nearest standard form. | Mark such passages and check them by ear. |
| Looking for real speaker names | Voices are labelled Speaker 1, Speaker 2 and so on. | Rename them in the exported document. |
More
See the full list of languages, the tools and the glossary.
Frequently asked questions
How good is Italian in published tests?
The large-v3 model has a published word error rate of 2.31 for Italian, which is among the lowest reported for the languages we list.
Are these your own measurements?
No. They come from tests published by others, and we did not run our own measurement for Italian.
Does regional Italian work?
The published figures do not split results by region, so try a short clip from your speaker first.
Will the text stay in Italian?
Yes. The transcript is written in the language of the recording and is not converted into another language.
Can I get subtitles for a video?
Yes. Switch timestamps on and export SRT or VTT.
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
- TheWhisper benchmark (turbo, average over FLEURS, MLS, Common Voice) · 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