Lip sync for video translation:
the lips move with the new speech
The translated video looks as if the person really speaks the other language. One button after the translation.
РусскийOne video, one voice clone, three languages. Without lip sync the mouth follows the Russian speech; with lip sync — the English, Chinese or Arabic one.
Original: «Откровения Продюсера», «Как сделать, чтобы твои видео выглядели дорого…» (youtube.com/watch?v=twytRncMLsU), CC BY 3.0. Translation: Speeek.
EnglishCuts between shots and an object in hand: lip sync works on the speaker, the phone inserts stay as they were.
Original: Tyron Tech, «Samsung Galaxy A05 Review», CC BY. Translation: Speeek.
РусскийShot on a phone with the face close to the camera — an easy case for lip sync. The English version for Shorts and Reels.
Original: Теплица социальных технологий, «Как правильно снимать видео на телефон», CC BY. Translation: Speeek.
Where lip sync works and where it is not worth trying
Lip sync redraws the speaker’s mouth, so it needs to see the face. We ran deliberately hard videos through it so you know in advance which of yours will come out best.
Works
The face is visible and the mouth is open to the model. Standard quality is enough.
Facing the camera, close-up or medium shot
Face to the camera, even light, nothing covering the mouth: courses, reviews, talking heads.
Cuts and objects in hand
Cuts do not get in the way: the lips are redrawn where the speaker is in the frame, inserts stay as they were.
Vertical phone video
The face is close and large — the model has plenty to work with.
Does not work
The model has nothing to hold on to. Better leave these videos without lip sync.
Wide shot, small face
The model could not find the mouth: you hear English while the lips follow the Russian. Same with a profile view — only one edge of the mouth is visible.
Hand, microphone or cup near the mouth
An object in front of the lips breaks the redraw: on Standard the mouth appears on top of the fingers.
Two people in the frame
Two half-turned small faces — the model found neither, the video stayed as it was. With larger faces it can go the other way: the mouth goes to the wrong person.
Standard or Pro
One translation, two models, both videos play in sync. While the face is clear there is almost no difference — it shows up from 0:26, when the hand is near the mouth.
Standard quality
For videos where nothing covers the mouth: talking heads, courses, reviews. On a clear face it is hard to tell from Pro.
Pro quality
When a hand or a microphone covers the mouth, for fast speech and close-ups: keeps the lip contour where Standard draws it over the fingers.
Three clicks after the translation
Lip sync is a separate step in the editor: first check the text and the voice, then sync the lips.
Open the video in the editor
Translation done and checked? Under the video on the right — the “Lip sync” tab, next to voices and sound.
Pick the quality and start
Standard or Pro — you see right away how many minutes it costs. Click “Make lip sync”: a few minutes later it is ready.
Compare and download
“Original · Translation · Lip sync” appears above the player — switch from the same spot. “Download” gets the lip-synced video.
Questions about lip sync
What exactly does lip sync do?
The mouth movements in the video are redrawn to match the new speech, so the person looks as if they really speak the translated language.
How much does lip sync cost?
Standard quality costs as many minutes as the video lasts: a 5-minute video — 5 minutes from your balance. Pro quality for close-ups costs 10× more. Lip sync is available on the Pro and Business plans and starts after the translation.
Do I need to translate the video again to add lip sync?
No. Lip sync runs on a finished translation — from the “Lip sync” tab in the editor. First check the text and the voice, then sync the lips; the translation stays, and above the player you can switch between “Original · Translation · Lip sync”.
Can I do lip sync without translation?
Yes: pick the same target language as the original — the video gets re-voiced with another voice or your clone, and lip sync matches the lips to the new track. That is how videos with bad audio get re-voiced.
Does it work on vertical videos and Shorts?
Yes, the format does not matter. The face does: large, in the frame and turned to the camera — which is usually the case in phone videos.
What happens if the video is not a good fit?
Most often the model simply cannot find the face — the video stays as it was: you hear the translation while the lips follow the original. That is what happened with the wide shot and the two half-turned faces in the gallery above. Less often it is the opposite: the mouth goes to the wrong person, or a photo on the wall comes alive. A run is charged either way, so check your video against the gallery first — for interviews, cut it so one person is in the frame at a time.
How the models move the lips and why they get lost in profile — in the article “How AI lip sync works”.
See it on your own video
Translation with a voice clone — 5 free minutes. Lip sync — on the Pro and Business plans.
Try a translation