Choose a voice and preview it
Pick from the available AI voices and listen to a sample before generating the dub.
AI dubbing & voice translation
Dub translated videos with AI voices
Translate a video's subtitles and then generate an AI voice track that speaks them in the target language. Choose a voice, preview it before you pay, and mix the dubbed audio into the final video — while keeping the option to export subtitles only.
How AI video dubbing works
After the video is transcribed and its subtitles are translated, the dubbing step sends each subtitle line to a neural text-to-speech engine. A single selected voice reads the full translated transcript, the voice track is mixed with the source video, and the result is rendered with the dubbed audio. The current product uses one voice per video; it does not yet separate speakers or match lip movements.
Key benefits
Subtitles or dubbing, your choice
Export only translated subtitles, or mix an AI voice track into the rendered video.
Keep the original audio
The source audio is never overwritten — the dubbed track is added as an optional layer.
How it works
- 1Upload a MP4, WebM and MOV video and translate its subtitles.
- 2Open the dubbing panel, preview the available voices and pick one.
- 3Generate the dubbed voice track for the translated subtitles.
- 4Render the video with the dubbed audio, or keep subtitles only.
Common use cases
- Localize explainer and marketing videos for international audiences.
- Make course and training videos watchable in another language.
- Add a translated voice track to interviews for listeners who prefer audio.
Frequently asked questions
Does dubbing replace the original audio?
No. The original audio is kept and the AI voice track is mixed in as an optional layer when you render the video.
How many voices are available and can I preview them?
Several voices are available across the supported languages. Every voice can be previewed before generation so you can hear it first.
Is this multi-speaker or lip-synced dubbing?
The current product uses one selected voice for the whole video and does not perform lip-sync. Speaker separation and lip movement matching are on the roadmap, not part of the current workflow.