AI Video Dubbing into Another Language
Online tool for translating and dubbing videos: it recognizes the speech, translates it to fit the timing and voices it in the language you pick. It tells up to three speakers apart, gives each of them a voice of their own and keeps the background music. The video is parsed and reassembled right in your browser, and every artifact downloads separately: the dubbed video, the audio track, subtitles and the translated script.
How to dub a video
Video or audio up to 60 minutes: MP4, WebM, MOV, MP3, WAV.
The source language is detected automatically; choose one of nine dubbing languages.
Basic in-browser voices or studio Premium voices — play a sample before you choose.
Fix the translation line by line, A/B against the original, download the video, track or subtitles.
Voice your video in another language with studio AI voices
Hear an example: one video, five languages — switch on the fly
⚠️ Processing runs in your browser: keep this page open and don't let the screen turn off.
Translation, line by line
The standalone track fits YouTube multi-language audio uploads.
🌍 Need another language version? Change the dubbing language above and run again — the speech is already recognized, so it is faster.
🤖 The translated voice-over is AI-generated.
Speech recognition, translation and voices are neural. Review the translation before publishing.
What it does
Typical uses
- An English version of your video for YouTube, with a standalone track for multi-language audio.
- Translating a course or tutorial into your audience's language with the background music intact.
- Localizing an ad for several markets without a dubbing studio.
- Voicing an interview or podcast with two or three speakers in another language.
- Bilingual SRT subtitles as a side product — for editing or review.
Why this one
This dub is not a robot reading a translation: lines are translated to match their screen time, each speaker gets a voice of their own, the music does not disappear, and the result ships in parts — video, track, subtitles, text. Edits never re-voice the whole video: only the changed line is rebuilt.
FAQ
Will the dubbed voice sound like the original speaker?
No — the dub is voiced with ready-made voices, not with a copy of the speaker. Pick one close to the speaker's gender and manner by ear: voices in the list come with a sample in the dubbing language. In a dialogue each speaker gets a voice of their own, so two people never merge into one narrator.
How does the tool tell the speakers apart?
By the sound of their voices: each detected speaker gets a separate dubbing voice. If it merges two people into one or splits one person in two, set the number of speakers by hand in “Speakers” before you start.
Translations run longer than the original — what about timing?
Translation is isometric: line length is fitted to its screen time at the translation stage, then the dub is gently distributed into pauses. Speech never turns into a chipmunk sprint.
What happens to the background music?
Enable "Keep music & background" — a neural model separates speech from the bed, and the music is laid back under the new dub without remixing.
Can I fix the translation?
Yes — the translation is editable line by line before and after voicing. An edit rebuilds only that line; nothing else is re-voiced.
What can I download?
The dubbed video (the picture is not re-encoded), a standalone WAV track, SRT subtitles in both languages and the translated script.
What happens to the tokens if a dub fails halfway?
Nothing is taken. The cost is held when the job starts and captured only when a finished voice-over is handed to you; a failure at any point returns the hold to your balance, and so does closing the tab. The hold is sized by the duration of the speech, which is why a long video shows a large number before anything has actually been spent — that number is reserved, not charged. If one line comes out wrong and the rest is good, re-render that line rather than the whole video: only the re-rendered part is billed, and it keeps the timing of everything around it intact.
Experts on this topic
Work through it with a specialist — an online lesson or consultation, times shown in your time zone.
We can! Just send us a quick message with your idea. If you'd like to discuss it in detail, leave your email and we'll get back to you. You can stay anonymous.
How do you rate this tool?
Comments
this dint translated from english to arabic with characters of male and female, male to male
Thank you — this was a real bug, and your wording ("male and female … male to male") is what made it findable. Two things were wrong. Voices were assigned by speaker number instead of by how the speaker actually sounds, so a man could be read by a female voice — and when the tool decided a clip had only one speaker, no voice was chosen at all. On top of that, voice pitch was judged by an average, and the pitch detector occasionally jumps an octave; those jumps pulled male voices into the female range, which is also why two people sometimes ended up merged into one speaker. Both are fixed and already live. Each speaker's voice is now chosen from that speaker's own pitch, including when there is only one speaker, and the panel shows what was detected. When a voice sits near the male/female boundary the tool says so instead of guessing, and you can switch that speaker's voice in one click — it applies to all of their lines. The English→Arabic translation itself was working; the problem was only in which voice read the result. Please give it another try and tell us how it sounds.
By uploading content you confirm that you are entitled to disclose its contents to third parties, and you undertake not to include other people's personal data in it. See the Terms of Use for details. AI providers Tembrica is an independent product. We are not affiliated with, endorsed by or sponsored by the makers of the models and services listed here; their names are used only to say factually which engine runs a given tool.
This result was generated by a neural network. Any rights that could arise on our side pass to you — perpetually, with no usage fees and no limits on territory, print run or manner of use, commercial use included. The rights are non-exclusive: another user may obtain a similar result. Crediting Tembrica or linking back is not required. The file carries a machine-readable AI-origin marker in its metadata. More in the Terms of Service