Caption French-language videos with correctly accented text, produced by a speech model running on your computer or phone. Nothing is uploaded, and there’s no account or watermark.
Open your French video or recording. The page starts with “Spoken language” set to French, so you can go straight on.
Choose “The spoken language” under “Subtitles in” for French captions, or “English (translated)” to subtitle a French video for English speakers.
Click “Generate subtitles” and watch the French lines arrive in the list while the percentage climbs.
Check proper nouns and figures, then save with “Download SRT” or “Download VTT”, or grab the full text with “Copy text”.
French is one of the languages Whisper handles best, because so much French audio was available to train it. Clear speech from a podcast, a class or a vlog usually comes out with correct grammar, accents (é, è, à, ç, ô) and apostrophes in elisions like l'école or qu'il, without any manual fixing.
Accents from Quebec, Belgium, Switzerland and West or North Africa are generally understood, though regional vocabulary, anglicisms in Quebec French and fast informal speech produce more errors than broadcast French from France. When several speakers interrupt each other, or music plays under the voice, expect to correct a few lines.
French typography puts a non-breaking space before ? ! : and ;. Subtitle files don't need it: SRT and VTT are plain text, and players wrap lines on their own. If your text contains those spaces, they are kept as written; if not, nothing breaks. What matters more is line length, since French sentences tend to run longer than their English equivalents.
This page is useful for creators publishing in French, teachers preparing accessible course videos, and students who want a written version of a recorded lecture to revise from.
Windows
Additional languages such as French arrived in a 2023 Windows 11 update, so older builds offer English only. The captions are not saved anywhere.
Android
French captions are available on some recent phones, notably Pixels, but not on all Android devices. Nothing is exported.
Keep French captions for a French-speaking audience, for accessibility, and for language learners who want to read what they hear. Accurate same-language captions also help viewers who watch with the sound off.
If you want to reach English speakers, the “English (translated)” option writes English subtitles straight from the French audio. It saves a step, but idioms can come out literally, so read it once before publishing. Whisper cannot do the reverse: it does not translate English speech into French.
For long French sentences, lowering “Max characters per line” to around 37 and keeping two lines can make subtitles easier to read on phones. Press “Re-split subtitles” after changing it.
Yes. The subtitles are written in UTF-8, so é, è, ê, à, ù, ç, œ and the apostrophe all survive the download and display correctly in modern players, on YouTube and in text editors.
Generally yes. Whisper heard many French accents during training. Very local slang, strong regional pronunciation or a mix of French and another language lead to more mistakes, which the “Best” setting reduces.
Not for subtitles. Players do not apply French typography rules to SRT or VTT, and a normal space or no space both display fine. Add them only if your style guide requires it.
No. The speech model only translates into English. For an English video, you can generate English subtitles and translate the text separately.
On a recent laptop with WebGPU, typically well under a minute once the model is downloaded. Without WebGPU, “Balanced” takes about as long as the video or a bit less. Phones are slower.
Yes. Because the recording is processed on your own device and never uploaded, it is a reasonable option for class recordings. Proofread the captions before sharing them with students, especially technical vocabulary.
Last updated October 28, 2026
Choosing the language is faster and more accurate than detection, especially for short clips.
Speech recognition makes mistakes, especially with names, numbers, music and several people talking at once. Dialects such as Moroccan Darija are recognised less reliably than Modern Standard Arabic. Read the subtitles before you publish them.
Need long videos or the best accuracy?
AI subtitles Pro will use Whisper large-v3 on our servers: about an hour of video in a few minutes, better accuracy on accents and dialects, files up to 3 hours. Only compressed audio is sent, never the video.
See Pro and join the waitlist