Local · Offline · 6 languages · Free & open source

Dub any short video like it was made that way.

Re-voice any clip into another language — with the speaker's own voice cloned, captions translated, and on-screen text localized right on the frame. 100% on your machine. No cloud, no upload, no subscription.

↓ Download for Windows ★ Star on GitHub
Windows 10/11 · NVIDIA GPU · one native .exe · zero Python
Dub Studio editor — dubbed video with translated captions and localized on-screen title

Before / after — real clips, dubbed offline

Every clip below was processed end-to-end on a local GPU. Different videos, different modes, different languages — nothing left the machine.

🎙️ Full dub · EN → RU
Original
Dubbed
Voice cloned into Russian · translated captions · localized on-screen title
🗣️ Voice-over · EN → ES
Original
Voice-over
Spanish voice-over the ducked original · 2 speakers, distinct voices
✨ Funny remix · EN → RU
Original
Remixed
The model rewrote the whole script to a new theme, then re-dubbed it
🎙️ Full dub · EN → ZH
Original
Dubbed
Cloned into Chinese · a real talking-head; baked-in English title blurred and re-localized
🎙️ Full dub · EN → PT
Original
Dubbed
Cloned into Portuguese · burned-in original captions cleaned and re-subtitled
🗣️ Voice-over · EN → FR
Original
Voice-over
French voice-over the ducked original · single speaker

Five modes, switchable on the fly

Load a clip once and send it into any mode — right inside the editor.

🎙️

Dub

Full re-voice into the target language with the original timbre cloned — auto-cast per speaker or pick a voice.

🗣️

Voice-over

Translated voice over the ducked original — the source is still audible underneath, balance adjustable.

📝

Subtitles

Burn original-language captions and keep the original audio — no dubbing, no translation.

Funny remix

Give a theme — the model rewrites the whole script, then re-dubs it.

🎬

Transcript

Clean diarized transcript with karaoke play-along, one-click voices, and .srt / .txt export.

Everything, running on your machine

🎙️

Voice cloning — the original timbre speaks the new language (Higgs Audio v3).

👥

Speaker diarization — a distinct voice per speaker (NVIDIA Sortformer, up to 4).

🖼️

On-screen text localization — OCR blurs the original and prints a matched localized title.

🌐

Local translation + vision — Gemma-4 12B translates and reads the frame layout.

🎚️

SOTA vocal separation — Mel-Band Roformer keeps the backing track intact.

💬

26 caption presets — karaoke, word-by-word, neon and more, WYSIWYG.

Live editor — ~0.17 s/frame preview; every change visible instantly.

📦

Batch processing — a queue of files with one setup.

🌍

6 languages — EN / RU / ZH / ES / PT / FR, source auto-detect.

🧳

Fully portable + auto-update — nothing touches your user profile.

How it works

A smart auto-pass builds the first draft; then you fine-tune everything with a live preview.

Drop a video and pick a target language — separation, ASR, diarization, translation and OCR run in one pass.

Choose a mode: dub, voice-over, subtitles, funny remix or transcript.

Edit anything — transcript, voices, caption style, blur boxes, titles — with an instant preview.

Export the finished video. Everything runs locally on your GPU.

Get Dub Studio

Free and open source. Download the portable build or the installer.