Finnish transcription is a good case for Potok. Published tests of the large-v3 model report a word error rate of 7.7 for Finnish, which is lower than for many larger languages. Send voice messages, audio or video to the Telegram bot; files up to 2 GB are accepted.
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 7.7 for Finnish comes from a published large-v3 test, not from our own runs.
- Finnish makes a meaning difference between short and long sounds, as in tuli and tuuli, so a doubled letter matters.
- Words carry many endings, so one error in an ending changes a whole word in the count.
- The transcript stays Finnish; the bot does not translate.
What a transcript looks like
[00:00] Speaker 1: Aloitetaan kokous. Ensin käymme läpi viime viikon tulokset ja sitten suunnittelemme seuraavat vaiheet.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 |
|---|---|---|---|
| Finnish (Suomi) | Recommended | word error rate 7.7% (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
- Open the chat first. Go to @potok_now_bot in Telegram.
- Share the recording. A voice message, forwarded clip, video or YouTube link works. Use Upload in the mini app when a file is close to the 2 GB limit.
- Look for doubled letters. Compare words with a long vowel or consonant against the audio; they are the most common single-letter errors.
- Check the endings of names. Finnish forms a name with several endings, so restore the base form before you file the text.
- Save. TXT or MD for plain text, DOCX or PDF for sharing, SRT or 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 |
|---|---|---|
| Overlooking a missing double letter | Tuli and tuuli mean different things, and spell-check may not catch it. | Read the key sentences aloud against the recording. |
| Splitting a long compound word | The sense is lost and search misses it. | Join compounds when you edit. |
| Quoting 7.7 as our own accuracy | The number describes another group's test set. | Say it comes from published tests. |
More
See the full list of languages, the tools and the glossary.
Frequently asked questions
How accurate is Finnish transcription?
Published tests of large-v3 give a word error rate of 7.7 for Finnish. Your own audio can score higher or lower.
Does Finnish count as a recommended language?
Yes, the published result puts it among the good languages on our list, without a warning.
Does it keep the letters ä and ö?
It should, but verify them. Losing the dots changes the word, so search the text for plain a and o where the dotted letters belong.
Can I add timestamps?
Yes, timestamps and speaker labels are optional and the labels come out as numbers.
Is any translation offered?
No. The text comes back in Finnish, and the bot does not translate transcripts into other languages.
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