Subtitle translator

Translate .srt and .vtt subtitle files between 40+ languages, with every cue timing kept exactly as it was. Plain text works too. Nothing is uploaded — the model runs entirely on your device, so it's free, unlimited, and private.

  1. Your text

    Drop a .srt, .vtt or .txt file here — or click to choose

    SRTVTTTXT

    Your files are never uploaded — everything runs on your device.

  2. Languages

  3. Translation

How to translate a subtitle file

  1. Add your file or text

    Drop a .srt, .vtt or .txt file, or paste text straight in. It never leaves your browser.

  2. Pick the languages

    Choose the language you're translating from and to. NLLB-200 covers 40+ common languages here, 200+ under the hood.

  3. Translate & download

    The model runs on your device (first run downloads it, then it's cached). Subtitle timings are preserved; download the translated file or copy the text.

Why translate here

Your files are never uploaded

Translation runs 100% in your browser — no server, no account, no upload. The perfect fit for confidential documents and unreleased subtitles.

Perfect subtitle timings

SRT and VTT cues keep their exact timestamps and order — only the spoken text is translated, so your subtitles still sync frame-for-frame.

Free & unlimited

Because inference happens on your device, there's no per-word cost and no cap. Translate as much as you like.

The model behind it

Powered by NLLB-200 (distilled, 600M), Meta's No Language Left Behind model, covering 40+ languages in the picker. It downloads once and then translates with no server round-trip at all.

Questions

Are my files uploaded to a server?
No. The translation model is downloaded to your browser and runs on your own device — your files never leave it.
Which formats can I translate?
Subtitle files (.srt and .vtt) and plain text (.txt), plus text you paste directly.
Which languages are supported?
40+ common languages are in the picker (English, Spanish, French, German, Chinese, Japanese, Arabic, Hindi and more), powered by Meta's NLLB-200 model.
Is it really free?
Yes — free and unlimited. There's no account and no metering because the model runs on your device, not our servers.
Why is the first translation slower?
The first run downloads the model (about 900 MB), then it's cached in your browser, so later translations start instantly.
Where does the translation happen?
Entirely in your browser. NLLB-200 distilled-600M is downloaded once, cached on your device, and run locally — so subtitles and documents you translate are never sent anywhere.