Free Tagalog transcription calls for a careful clip test, because the published word error rate on the earlier large-v2 model is 13.8, close to the 15 line we treat as borderline. These are tests of others, and Potok had no large-v3 figure. Read a short sample 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: 13.8 is near our 15 line, so Tagalog is borderline and the text needs a human read.
- No large-v3 figure was found, so the number is from large-v2 and from other people's benchmark, not ours.
- Everyday Tagalog often switches into English mid-sentence, so a good sample includes that mix.
- Words are often built with affixes and repeated syllables, which can confuse spelling.
What a transcript looks like
[00:00] Speaker 1: Simulan na natin ang pulong. Una, susuriin natin ang mga resulta noong nakaraang linggo, pagkatapos ay magpaplano tayo ng mga susunod na hakbang.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 |
|---|---|---|---|
| Tagalog | Recommended, borderline | word error rate 13.8% (large-v2, FLEURS) Whisper paper | Published tests show an error rate close to our 15% line. Check a short sample first. No large-v3 figure was found; the figure is from the earlier large-v2 model and sits close to the 15% line. |
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 real, mixed speech. Take a minute of how people actually talk, including the English words they drop into a sentence.
- Open the bot. Open @potok_now_bot in Telegram and send the clip as a voice message, an audio file or a video.
- Mark the switch points. Look at the exact places where the speaker moves between languages and see which words went wrong.
- Check repeated syllables. Compare words with doubled parts, such as plurals or intensifiers, with the audio.
- Plan the proofread. If the sample is usable after edits, 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 |
|---|---|---|
| Expecting a clean result from borderline numbers | A figure this near the line means a visible share of mistakes. | Budget time for a full read-through. |
| Testing only pure Tagalog | Real talk is mixed, so a clean test flatters the tool. | Add a clip with English words in it. |
| Skipping the check on names | Personal names are guessed from sound. | Keep a list of names to compare. |
| Reading the text without audio | Plausible-looking errors slip through. | Play the clip while reading. |
More
See the full list of languages, the tools and the glossary.
Frequently asked questions
Why is Tagalog marked with a caution?
The published error rate is close to our borderline, so careful checking is part of the job.
Which model does 13.8 refer to?
The earlier large-v2 model, because no large-v3 figure was found.
Is the number from your own tests?
No. It comes from published tests of others.
Can I get English text out of it?
No. The transcript stays in the language of the recording.
What is the file size limit?
Files up to 2 GB are accepted, and a long result arrives as a file you can save.
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
- Whisper paper, Table 13 (large-v2 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