Croatian is on the list of good results for Potok, which transcribes it through a Telegram bot. In published tests of the large-v3 model the word error rate for Croatian is 11.07. The bot takes voice messages, audio, video and files 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
- The word error rate of 11.07 for Croatian is a published large-v3 result, not a test we ran.
- Croatian uses the Latin alphabet with the letters č, ć, đ, š and ž, and a dropped mark can change a word.
- Nouns and names take case endings, so a surname may appear in a different form from the one you expect.
- Remember that any score describes tested speech; phone calls and crosstalk usually do worse.
What a transcript looks like
[00:00] Speaker 1: Počnimo sastanak. Prvo ćemo pregledati rezultate prošlog tjedna, a zatim isplanirati sljedeće korake.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 |
|---|---|---|---|
| Croatian (Hrvatski) | Recommended | word error rate 11.07% (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
- Send it over. Write to @potok_now_bot and attach a Croatian voice message, audio file, video or YouTube link.
- Choose a route for size. A normal attachment works up to 2 GB; when Telegram is slow with a big file, use the Upload button.
- Scan the diacritics. Search the text for words with č, ć, đ, š and ž, since a plain c, s or z in their place is a typical slip.
- Fix endings of names. Compare surnames with the nominative form used in your document.
- Save the result. TXT and MD for notes, DOCX and PDF for sharing, SRT and VTT for subtitles.
| 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 |
|---|---|---|
| Accepting letters without their marks | Search and spell-check fail on words such as ćevapi if the mark is lost. | Review and restore the marks in a word processor. |
| Copying declined names straight into a form | The name appears in the wrong case. | Change the ending back to the base form. |
| Expecting speaker names | Labels come out as numbers, not people. | Rename Speaker 1 and Speaker 2 in the exported file. |
| Using a muffled call recording | Quiet endings of words disappear first. | Send the original file instead of a re-recording. |
More
See the full list of languages, the tools and the glossary.
Frequently asked questions
How accurate is Croatian transcription?
Published tests of large-v3 show a word error rate of 11.07 for Croatian. It is another group's number, so check your own audio.
Will the Croatian special letters appear?
Usually yes, but check them. Letters such as č, ć, đ, š and ž matter for meaning, so scan for them before use.
Is Croatian recommended or cautioned?
It sits among the good results on our list, without a warning from the published data.
Can I upload a long lecture?
Yes, up to 2 GB per file, and the Upload button in the bot's mini app is made for big files.
Does the bot translate the text?
No. The transcript stays in Croatian and no translation is offered.
Sources
Each source was opened by hand on the date shown. Pages change, so open the source if a detail matters to you.
- FLEURS speech dataset (Google, Hugging Face), Hugging Face · checked
- Word error rate (Wikipedia), Wikipedia · 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