Middle East and North Africa · Arabic · العربية
Arabic subtitle generator: audio to SRT with local Whisper
This Arabic subtitle generator runs Whisper locally in your browser, so Arabic audio stays on your device while you build an SRT or VTT file. It drafts Arabic script, not a promised romanized version, and every right-to-left cue still needs a human review before publishing.
By Abdul Samad · Last updated
Generate subtitles
Runs Whisper in your browser. No upload, no account, no credit card.
Choose a video or audio file
Drop it here, or click to browse. MP4, MOV, WEBM, MP3, WAV, M4A and similar formats.
Your file never leaves your device
Up to 500 MB. Files over 100 MB can be slow on phones.
Only transcribe media you own or have permission to use. VoiceToTextFree does not download media from other sites.
Auto-detect works for clear speech; choosing the language is faster and safer for accents.
Limits: 500 MB per file; speech must be audible; Whisper can struggle with music, overlapping voices and very noisy recordings. VoiceToTextFree does not upload files or send audio to a paid API.
Key facts
| Item | Detail |
|---|---|
| Language code | ar |
| Native name | العربية |
| Script | Arabic, usually set in Naskh |
| Direction | Right to left |
| Region | Middle East and North Africa |
| Common digits | Arabic-Indic ٠١٢٣٤٥٦٧٨٩ or Western 0123456789 |
| Output | SRT, VTT and TXT |
Start with an Arabic subtitle generator that keeps audio local
An Arabic subtitle generator has to handle a language with a long written tradition and a very wide range of spoken forms. VoiceToTextFree runs a Whisper model in the browser, so the audio is processed on your machine rather than uploaded to a transcription service. That helps with confidential meetings, unpublished interviews and family recordings. You select Arabic in the picker, let the model draft the cues, then review the text before exporting SRT, VTT or plain text.
The draft is not a finished subtitle track. Whisper usually returns Arabic in Arabic script, and it does not promise a romanized transliteration for mixed or unclear speech. Expect to correct spelling, diacritics, punctuation and line breaks. The review pass is where most of the quality comes from, and it is where you catch the names and numbers a model cannot know.
Arabic audio to SRT: what the file actually contains
When you export Arabic audio to SRT, you get numbered cues with start and end timestamps and the Arabic text saved as UTF-8. The file has no direction field. The direction is not stored in the file: the player applies Unicode's bidirectional rules to the characters. Because different players interpret those runs differently, keep the logical character order and never mirror a line manually.
Cue length is a reading constraint, not a speech constraint. Split a long sentence into two or three cues when the viewer needs more time. The exporter writes exactly what you approve, so correct the wording in the editor instead of patching the SRT later in a program that might rewrite its encoding.
Arabic script, diacritics and punctuation
Arabic is written in a cursive script in which letters join differently depending on their position in a word. Most published Arabic omits short vowel marks, so the reader supplies them from context. A transcription model may add diacritics inconsistently, and a word with different vowel marks can look unusual to a reader even when the consonants are correct. Review the final text rather than trusting every mark.
Arabic punctuation includes the comma (،), the question mark (؟) and the full stop. Modern text also uses Western punctuation, especially in technical and social media content. Choose one convention and apply it consistently. Avoid letting a word processor convert the marks, because smart punctuation and direction marks can change how a cue is displayed.
Right-to-left layout and Arabic captions
Arabic captions read from right to left, but the characters are stored in logical order. The player applies the Unicode bidirectional algorithm to display them correctly. If you reverse the text by hand, the player reverses it again and the line looks wrong in a different way. Encoding comes first: save as UTF-8 so the script survives the export.
Alignment is separate from direction. Centred captions work for many mixed-script videos, while start and end alignment should follow the script rather than a fixed left or right edge. A left-aligned Arabic cue is not automatically broken, but it often looks disconnected from the reading flow. Test the rendered cue in the player before deciding.
Arabic subtitles for YouTube and social video
For Arabic subtitles for YouTube, the caption box is often narrow and many viewers watch on phones. Favour brief lines, cap a cue at two lines, and choose a syntactic pause instead of a character count. Joined Arabic letters can make a short-looking line feel heavy once the player sets it.
Watch the captions at normal playback speed. When a cue vanishes before the sentence ends, either divide the cue or tighten the wording. Put the file on an unlisted copy first and inspect the rendered captions, because the platform uses its own font stack.
Mixed Arabic and English speech
Many Arabic speakers switch between Arabic and English or French inside a sentence, and technical content is full of Latin brand names. Whisper tends to output the script of the language it detects, so a heavily code-switched clip can produce a mixture of Arabic and Latin words. That is normal, but it needs review: the model may write a foreign term in Arabic letters or leave an Arabic word in Latin spelling.
We do not promise romanized output. If you need Latin transliteration, treat it as a separate editorial task and do it deliberately rather than assuming the model will. For captions, keep the script your audience expects and make the mixed-language lines consistent across the whole file.
Names, numbers and Arabic-Indic digits
Arabic text may use Arabic-Indic digits (٠١٢٣٤٥٦٧٨٩) or Western digits (0123456789), and both appear in real broadcasts. Pick one digit set for the audience and use it consistently. Digits can reorder visually beside right-to-left text, so test a phone number, a date range or a room number in the target player.
Proper nouns need particular care. The output may use a phonetic spelling or another plausible variant. Ask the speaker or check an official spelling, and maintain a short glossary of recurring names, organisations and technical terms for the review pass.
Arabic audio to text: dialects, accents and standard Arabic
Arabic is a family of spoken varieties. Broadcasters often use Modern Standard Arabic, while interviews and social video may use Egyptian, Levantine, Gulf, Maghrebi or another regional dialect. A model trained on broad data may handle a clear standard register better than a fast dialectal conversation, and the written output may drift toward standard spelling. That is a normal limitation, not a sign that the audio is unusable.
Punctuation is inferred from pauses, and the inference can misrepresent the speaker. Rewrite a cue when a mark changes the sense, and drop marks where there was no pause. Overlapping speakers are difficult; separate the audio tracks or accept gaps in the transcript.
Review workflow and privacy for Arabic files
Review in three passes: listen with the draft open, fix script and names, then watch the export in the target player. Inspect the opening cue, the longest cue, a numbered cue and one with Latin text mixed in. Check that the export is UTF-8 before publishing.
Transcription happens in the browser, so this step does not send the audio off the device. That lowers exposure for private meetings, client calls and unreleased content. Consent is still required, and an unchecked transcript is not ready to publish. Keep the local model as a fast first draft and retain human review.
Illustrative Arabic cues with an English translation note
This short sample shows how an Arabic draft might look in an SRT file. It is illustrative only, not a benchmark or a measured accuracy result.
1
00:00:01,000 --> 00:00:04,000
الاجتماع يبدأ الساعة التاسعة صباحًا.
2
00:00:04,200 --> 00:00:08,000
يرجى إحضار التقرير الأزرق إلى الغرفة 402.Frequently asked questions
Does this Arabic subtitle generator upload my audio?
No. The model runs in the browser for this workflow, so the audio is processed on your device rather than sent to a transcription server. You should still check your browser and device settings, and get consent before transcribing someone else's private recording.
Does Whisper output Arabic in Latin letters?
Whisper normally returns Arabic in Arabic script. It does not promise a romanized transliteration, and code-switched speech can produce a mixture of scripts. If you need Latin transliteration, treat it as a separate editorial step.
Can I upload an Arabic SRT to YouTube?
Yes. Export SRT or VTT, then upload it as a caption track on a private copy of the video first. YouTube uses its own font stack and may reflow the text, so check the rendered captions on desktop and on a phone before publishing.
Why is my Arabic punctuation on the wrong side?
Punctuation can attach to the wrong directional run when a cue mixes Arabic with Latin text or numbers. Place the mark at the logical end of the sentence, keep numbers in one run and test the exact cue in the player. Do not reverse the characters by hand.
Does it work with Arabic dialects?
The model can transcribe dialectal speech, but results vary with the variety, accent, speed and audio quality. Modern Standard Arabic is usually the most predictable register. Review dialectal cues carefully because the written form may drift toward standard spelling.
How accurate is Arabic transcription?
We have not recorded a reviewed Arabic sample, so this page does not claim a tested accuracy level. Recognition varies with dialect, accent, audio quality and model size. Treat every draft as a first pass that still needs a human review.
Sources and further reading
Abdul Samad — Creator and maintainer. About Abdul Samad.
Last updated · Published
Related guides
- How to improve Whisper accuracy for cleaner subtitlesHow to improve Whisper accuracy with better audio capture, explicit language selection, measured model sizes, overlap handling and a practical human review workflow.
- Right to left subtitles and non-Latin scripts: a practical guideRight to left subtitles need UTF-8, bidirectional text, correct punctuation, suitable fonts and player testing. A practical guide to RTL SRT and WebVTT.
Other languages
Languages in the picker: 99. The list is checked against the Whisper tokenizer by a unit test.