Arabic speech to text, free and private
Drop an Arabic recording or video. Whisper runs on your device and writes an editable Arabic transcript with timestamps and subtitles.
Public-domain LibriVox recording. Load it, then press Transcribe to try the full tool.
Whisper base, the model this tool runs, scored a 48.8% word error rate on Arabic 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.
To transcribe Arabic for free, drop the file above; Whisper base runs in your browser and nothing is uploaded. The paper measures a 48.8% word error rate for this model on Arabic read speech (FLEURS) and 84.4% on the more varied Common Voice set, so treat the output as a rough draft and expect dialects to do worse.
Checked by Koldflux · updated 2026-09-25 · tested on a public-domain Arabic recording
How accurate is Arabic transcription?
Word error rate from the Whisper paper (lower is better): FLEURS, read sentences recorded under the ar_eg (Arabic, Egypt) 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) | 48.8% | 84.4% | Rough draft only |
| Whisper small | 30.6% | 66.4% | Rough draft only |
| Whisper medium | 20.4% | 60.3% | Usable, needs a proofread |
| Whisper large-v2 | 16% | 53.8% | Usable, needs a proofread |
What we saw on a real Arabic recording
We ran the first 30 seconds of a Modern Standard Arabic reading of al-Kawakibi’s book (LibriVox, public domain) through the same files this page loads, with the language set to Arabic. One 30-second clip is an illustration, not a benchmark. Processing took 8 seconds on one CPU thread of our test server.
الفصل صفر من طبعا عن الاستبداد ومصارع الاستعلاد هذا تسجيل خاص باليبريفوكس جميع تسجيلات لبريفوكس في المجال العام لمزيد من المعلومات أو للمشاركة زور الموقع لبريفوكس.org …- Detected as Arabic with 99.2% confidence.
- The LibriVox announcement (“all LibriVox recordings are in the public domain…”) came out essentially right.
- The book title was misheard: طبائع became طبعا and الاستعباد became الاستعلاد. Rare words and names are where it slips.
Modern Standard Arabic, dialects and right-to-left text
Whisper has one Arabic setting and writes in Arabic script, usually without vowel marks. The paper’s two Arabic tests disagree sharply (48.8% on FLEURS, 84.4% on Common Voice), which is a warning that results swing with speaker, accent and microphone.
Colloquial dialects (Egyptian, Levantine, Gulf, Maghrebi) are written in many different ways and are further from formal read speech; expect more errors than on Modern Standard Arabic.
SRT and VTT files store the text as plain Unicode, and players lay Arabic out right to left on their own. Check punctuation at line ends in your player before publishing.
Arabic subtitles and translation
| Value | |
|---|---|
| Characters per line (Netflix) | 42 characters per line |
| Reading speed (Netflix) | up to 20 characters per second (adult programs) |
| Tool default line length | 42 characters |
| Speech to English, Whisper base (BLEU, higher is better) | 1 |
| Speech to English, Whisper large-v2 (BLEU) | 25.5 |
Tips for Arabic
- Arabic to English: transcribe firstWhisper base scores a BLEU of 1.0 translating Arabic speech to English (large-v2: 25.5). Transcribe in Arabic here, correct it, and translate the text instead.
- Arabic subtitlesNetflix’s Arabic guide allows 42 characters per line and about 20 characters per second for adults.
- Pick Arabic by hand for mixed speechIf speakers switch between Arabic and English or French, set Arabic manually so auto-detect does not flip whole sections.
Limits and when to use something else
- Dialect-heavy speech, poor phone audio and fast conversation will be well below the benchmark.
- The built-in English translation is not usable for Arabic with this small model.
- 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
Can this translate Arabic audio to English?
It has a translate option, but Whisper base scores a BLEU of 1.0 on Arabic-to-English in the paper, so the result is not usable. Transcribe in Arabic and translate the text.
Does it understand Egyptian or Gulf dialects?
It uses one Arabic setting for everything. It was measured on read Arabic (FLEURS, recorded under the Egypt locale) and on Common Voice; conversational dialect speech usually does worse.
Will the subtitles show right to left?
Yes in standard players: the files hold plain Unicode Arabic and the player applies right-to-left layout. Check line-end punctuation in your player.
How accurate is it?
Whisper base: 48.8% word errors on FLEURS, 84.4% on Common Voice. Whisper large-v2: 16.0% and 53.8%.
Turn one idea into a week of posts
Koldflux Studio (paid, from $19/month) transcribes Arabic uploads on our servers with Whisper large-v3-turbo (the paper measures 16.0% for the comparable large-v2 on FLEURS), then turns the recording into Shorts and carousel posts you review, approve and export. It does not translate. 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.