Free Korean transcription needs a careful test, because the published word error rate for large-v3, which Potok uses, is 14.1, close to the 15 line we treat as borderline. Those are tests of others, and spacing changes the count. Try a short clip and read it against the audio.
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
- Caution: 14.1 is close to our 15 line, so Korean is borderline and a manual read is part of the job.
- The figure comes from published tests of others on large-v3, not from our measurements.
- The count depends on how spaces are treated, so two tests of the same audio can disagree.
- Korean particles and endings attach to words, which makes a one-letter slip look larger.
What a transcript looks like
[00:00] Speaker 1: 회의를 시작하겠습니다. 먼저 지난주 결과를 검토하고, 그다음 다음 단계를 계획하겠습니다.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 |
|---|---|---|---|
| Korean (한국어) | Recommended, borderline | word error rate 14.1% (large-v3, FLEURS) Whale paper | Published tests show an error rate close to our 15% line. Check a short sample first. This figure is close to the 15% line we use, and it depends on how spaces are counted, so treat it as borderline. |
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
- Record or cut a short clip. Take about a minute with at least one sentence in polite style and one in casual style.
- Open the bot. Open @potok_now_bot in Telegram and send the clip as a voice message, an audio file or a video.
- Mark spacing separately. Wrong word breaks are common in Korean text. Tick them in a different colour from real hearing errors.
- Check endings and particles. Read the end of each phrase twice, since this is where the meaning shifts.
- Decide on long material. If the clip is readable after your fixes, send the full file, up to 2 GB, and export TXT, DOCX or SRT.
| 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 |
|---|---|---|
| Calling every spacing difference an error | Space rules are flexible enough that your count turns out inflated. | Tally spacing apart from wrong words. |
| Trusting the number without a clip of your own | A borderline figure can move either way on your audio. | Use the published value as a hint, then test. |
| Skipping the sentence endings | The ending carries politeness and tense, and a wrong one reads oddly. | Proofread the last syllables of each line. |
More
See the full list of languages, the tools and the glossary.
Frequently asked questions
Is Korean a good language for this?
It counts as good, but borderline, so plan for a proofread rather than a straight copy.
Where does 14.1 come from?
From published tests of others on large-v3, not from our own measurements.
Why does spacing matter?
Korean word breaks are flexible, so test scores change with how spaces are counted.
Can I get the text in English?
No. The transcript stays in the language of the recording.
What file size can I send?
Files up to 2 GB are accepted, so a long lecture or interview can follow once the short test looks right.
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
- Whisper speech recognition system (Wikipedia), Wikipedia · checked
- Whale paper, Table 5 (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