Translate the speech, keep the speaker's voice and re-sync the lips. Review every line before it's spoken.
Add the video to translate
MP4, MOV or WebM up to 200 MB · one clear speaker works best
Music and effects: Demucs. Music model: some sound effects may be lowered, and singing may be replaced.
Cloud: ElevenLabs isolates the voice and Demucs lifts out the music and effects → Scribe v2 times the voice → Qwen3.5 on your GPU translates → you review → ElevenLabs speaks each line at its time and Chatterbox turns it into the speaker's voice → Sync Lipsync 3 → mixed back over the music and effects.
Add the video to translate