AI lip sync
Make any video say what you type
Upload a clip of one person, write the script and pick a voice. The AI speaks it and redraws the mouth to match, so the result looks like it was filmed that way.
- 17 ready-made AI voices
- 6 languages, more via multilingual voices
- Works on your own footage
- No subscription
Real result, with sound
See more examplesVoice
Press play to hear a sampleWrite the script in this voice's language for the most natural result.
2 credits per second of speech at any resolution. The price shows as you type.
Sign up and get 60 free credits.
Example
One silent clip, any script
The clip on the left has no sound. On the right it speaks a typed script in an AI voice, with the mouth redrawn to match.
Turn the sound on. Only the mouth area changes; the face, background and framing stay as filmed.
How it works
From script to talking video in minutes
- 01
Upload a clip
Any MP4 up to 10 minutes where one person faces the camera. A few seconds is enough.
- 02
Type the script and pick a voice
Choose from 17 preset voices and play a sample of each. No recording or voice cloning.
- 03
Download the video
The AI speaks the script and moves the lips to match. Download an MP4 with the new voice.
Voices
17 natural AI voices
Warm narrators, clear presenters and friendly hosts, female and male, in six languages.
Chinese
Vivi (Female) · Xiaohe (Female) · Yunzhou (Male) · Xiaotian (Male)
English
Allison (Female) · Jenny (Female) · Alex (Male) · Jamie (Male) · Russell (Male)
Japanese
Hana (Female) · Ken (Male)
Korean
Momo (Female) · Jay (Male)
Spanish
Romina (Female) · Mariano (Male)
Brazilian Portuguese
Sofia (Female) · Arthur (Male)
The Chinese voices also speak English and 30+ other languages, detected from the script.
Tips
What makes a good clip
The lip sync redraws the mouth of the person in your clip, so it needs to see it clearly.
- One real personOne face in the frame works best. Cartoons, masks and crowds aren't supported.
- Facing the cameraLooking at the camera or slightly to the side (up to about 45°), head not tilted far up or down.
- Mouth visibleKeep hands, microphones and hair away from the mouth. Closed or gently moving lips sync best.
- Even light, SDRA sharp, well-lit face gives the cleanest result. Export HDR footage as standard MP4 first.
Use cases
Film once, say anything
Update what a video says without booking another shoot.
Ad variations
Test different hooks and offers on the same take, then keep the one that converts.
Product videos
Let a presenter introduce every product in your catalogue from one recording.
Courses and training
Fix a line or update a lesson without filming the teacher again.
Creator content
Make talking-head posts from a short clip when you can't be on camera.
Other languages
Have the same person speak Chinese, English, Japanese, Korean, Spanish or Portuguese.
Announcements
Turn release notes, offers and news into short spokesperson videos.
Pricing
Pay per second of speech
2 credits per second at any resolution. The price is set by the script and shown before you start.
| 15-second video | 30 credits |
|---|---|
| 30-second video | 60 credits |
| 60-second video | 120 credits |
| 120-second video | 240 credits |
- Same price for 720p and 4K
- No subscription, credits never expire
- Automatic refund if a job fails
FAQ
AI lip sync questions
It makes the person in your clip say a new script. The script is spoken by an AI voice and the mouth in every frame is redrawn to match the speech, while the face, background and camera stay as filmed.
An MP4 of up to 10 minutes and 2 GB with one real person looking at the camera, face turned no more than about 45° to the side. HDR footage isn't supported. A few seconds of footage is enough.
No. Pick one of the 17 preset AI voices and play a sample before you start. Nothing has to be recorded.
There are voices for Chinese, English, Japanese, Korean, Spanish and Portuguese. The Chinese voices can also read English and more than 30 other languages, including Hindi, Thai, Arabic, French and German, detected from the script.
The video is as long as the speech, up to about two minutes per script. The price and the expected length update as you type.
The clip plays forward and then backward until the speech ends, so a short clip still covers a longer script. If the clip is longer, the video ends with the speech.
Usually 3 to 5 minutes for a 15-second video, a little longer for longer scripts. You can close the page; the video appears under Recent videos when it's done.
2 credits per second of speech, whatever the resolution. A 30-second talking video costs 60 credits. New accounts get free credits to try it.
Not yet. For now the script is spoken by one of the preset voices.
Clips are only used to make your video. Results are kept for 5 days so you can download them, then deleted automatically. Only use footage of people who agreed to it.
Failed jobs are retried automatically. If the video still can't be made, the credits go back to your balance.
Give your video a new script
Upload a clip, type what it should say and download the talking video in minutes.