Tamil speech to text, free in your browser
The tool is set to Tamil. Drop a recording or video: the model runs on your device and gives you a rough Tamil draft with timestamps to correct.
Public-domain LibriVox recording. Load it, then press Transcribe to try the full tool.
Whisper base, the model this tool runs, scored a 58.7% word error rate on Tamil in the Whisper paper's FLEURS test (lower is better). Expect a rough draft: many words will need fixing.
First use downloads the speech model and engine (about 97 MB, of which 72.5 MB is the model); later visits load it from your browser cache. Runs on your CPU; nothing is uploaded. Up to 100 MB and 2 hours per file; under 30 minutes is the comfortable range on a laptop.
You can transcribe Tamil for free in the tool above, but expect a rough draft. Whisper base, which runs in your browser, scores 58.7% word errors on Tamil in the Whisper paper (FLEURS). Keep the language set to Tamil: in our test, auto-detect guessed Malayalam. Larger Whisper models do much better (large-v2: 17.5%).
Checked by Koldflux · updated 2026-09-25 · tested on a public-domain Tamil recording
How accurate is Tamil transcription?
Word error rate from the Whisper paper (lower is better): FLEURS, read sentences recorded under the ta_in (India) locale, and Common Voice 9, volunteers with varied microphones. This tool runs Whisper base. Koldflux Studio runs Whisper large-v3-turbo, which the paper does not measure; OpenAI reports large-v3 cuts errors by 10 to 20% against large-v2, and the turbo version trades a little of that accuracy for speed.
| Model | FLEURS WER | Common Voice WER | Verdict (FLEURS) |
|---|---|---|---|
| Whisper base (this tool) | 58.7% | 49.5% | Rough draft only |
| Whisper small | 35.2% | 28.7% | Rough draft only |
| Whisper medium | 23.1% | 19.6% | Usable, needs a proofread |
| Whisper large-v2 | 17.5% | 16.1% | Usable, needs a proofread |
What we saw on a real Tamil recording
We ran the first 30 seconds of a Tamil reading of the Tiruppavai (LibriVox, public domain) through the same files this page loads, with the language set to Tamil. One 30-second clip is an illustration, not a benchmark. Processing took 12.6 seconds on one CPU thread of our test server.
முதல் پگுதி, திருப்பாவை, பாசுரங்கள் ஒன்று முதல் 5 வரை. இது ஒருல்லி விற்வாகுச் சொல்லிப்பதிவை, அனைத்தில் விற்வாகுச் சொல்லிப்பதிவைகளும் பவிலிக் கடுமேனில் உள்ளனா …- Auto-detect was unsure and wrong: Malayalam 33.6%, Tamil 27.4%, Telugu 27.2%. With Tamil chosen, the output was in Tamil script.
- The opening “முதல் பகுதி, திருப்பாவை, பாசுரங்கள் ஒன்று முதல் 5 வரை” was nearly right, apart from a stray Urdu letter inside பகுதி.
- The English words in the LibriVox announcement (“LibriVox”, “public domain”) were written as Tamil-sounding nonsense.
Tamil script, long words and English words
Tamil joins suffixes onto words, so one misheard ending makes a whole long word count as an error; that is part of why word error rates look high even when the draft is readable.
English words said inside Tamil speech (Tanglish) tend to be written in Tamil script by sound, often wrongly. If a passage is mostly English, run that part again with English selected.
Tamil subtitles and translation
| Value | |
|---|---|
| Characters per line (Netflix) | 42 characters per line |
| Reading speed (Netflix) | up to 22 characters per second (adult programs) |
| Tool default line length | 42 characters |
| Speech to English, Whisper base (BLEU, higher is better) | 0.4 |
| Speech to English, Whisper large-v2 (BLEU) | 9.2 |
Tips for Tamil
- Always pick Tamil by handAuto-detect confused Tamil with Malayalam and Telugu in our test. This page presets Tamil; keep it.
- Correct, then exportFix the text in the editor before downloading: edits go into the SRT, VTT and TXT files.
- Tamil subtitlesNetflix’s Tamil guide allows 42 characters per line and about 22 characters per second for adults.
Limits and when to use something else
- Expect to rewrite a large share of words; this is a starting draft, not a finished transcript.
- Tamil-to-English translation with this model is not usable (BLEU 0.4; large-v2 reaches 9.2).
- Speed depends on your device. A laptop handles a 30-minute file comfortably; phones may run out of memory on long files. Hard limits: 100 MB and 2 hours per file.
- No speaker labels: Whisper writes one stream of text. It also does not mark music, laughter or background sounds reliably.
Frequently asked questions
How accurate is free Tamil speech to text?
Whisper base scores 58.7% word errors on Tamil FLEURS and 49.5% on Common Voice. Whisper large-v2 reaches 17.5% and 16.1%. Here that means a draft you correct line by line.
Why did it detect Malayalam?
Language detection is weak for closely related South Indian languages with this small model. Choose Tamil manually; this page does that for you.
Can it translate Tamil audio to English?
The option exists but the result is not usable with this model (BLEU 0.4 in the paper).
Is my audio uploaded?
No. Everything runs in your browser tab.
Turn one idea into a week of posts
For Tamil, a larger model roughly triples accuracy: the paper measures 58.7% word errors for base and 17.5% for large-v2. Koldflux Studio (paid, from $19/month) transcribes uploads on our servers with Whisper large-v3-turbo, then turns the recording into Shorts and carousel posts you review, approve and export. It does not publish or schedule for you.
- 1 · Recording, idea or script
- 2 · Pick the angles
- 3 · Review and approve
- 4 · Export Shorts and carousels
Language pages
Each page has the tool preset to that language, measured accuracy, our own test on a real recording and subtitle rules.