Home/ AI Lip Sync Video Generator
AI Lip Sync Video Generator
Give it a face and a voice. LTX-2 Lipsync matches the mouth in your image to an audio file you upload — the clip is exactly as long as the speech.
Lip sync puts speech onto a face. You upload an MP3 or WAV of what should be said and a reference image of who says it; LTX-2 Lipsync animates the mouth to follow the audio. Length is set by your audio, not a duration picker — up to about 20 seconds per run — and output renders at 480p, 720p or 1080p, priced per 5 seconds of audio.
- Model
- LTX-2 Lipsync
- Starts with
- An image and audio
- Output
- MP4
- Clip length
- Matches your audio length
- Resolution
- 480P / 720P / 1080P
- Sound
- Follows your audio
- Credits
- from 32 credits
- Commercial use
- Included — see pricing
Your audio drives the mouth
Upload an MP3 or WAV — speech, vocals, a voiceover take — and the clip length follows the file, so there is no dead air or cut-off sentence.
One face reference
A front-facing image with a visible mouth is all the model needs to decide who is speaking.
You see the length before you pay
The studio reads your audio duration and estimates credits before generation, so a 20-second take never surprises you.
Draft cheap, finish at 1080p
Preview at 480p while you tune the take, then render the final at 720p or 1080p.
How to sync a face to audio
- Step 1Upload the audioMP3 or WAV. Clean, single-speaker recordings sync best — heavy music under speech confuses the mouth.
- Step 2Add the faceA frontal image where the mouth is unobstructed. Sunglasses are fine; a hand over the mouth is not.
- Step 3Check the detected lengthThe studio shows the audio duration it read — that is your clip length and the basis of the credit estimate.
- Step 4Replay it against the audioWatch the consonants on the preview before you render the same take at a higher resolution.
Where lip sync pays off
Dub a character
Give an illustrated or generated figure real dialogue without rigging or a 3D pipeline.
Localize a talking clip
Swap the language track and regenerate the mouth instead of re-filming.
Voice memes and singing stills
A portrait that sings an audio clip is a proven social format.
Fix dialogue after the fact
Rewrite a line, regenerate the sync — cheaper than re-animating the whole shot.
Know before you generate
- Two inputs are required: an MP3/WAV audio file and a face image.
- Clip length equals your audio length, up to roughly 20 seconds per run.
- Credits are charged per 5 seconds of audio: 32 at 480p, 42 at 720p, 64 at 1080p.
- Clean single-speaker audio syncs noticeably better than noisy or multi-voice recordings.
Lip sync questions
Which audio formats work?
MP3 and WAV. The studio reads the duration when you upload and prices the run from it.
How much does it cost?
Per 5 seconds of audio: 32 credits at 480p, 42 at 720p, 64 at 1080p. A 10-second 720p clip is about 84 credits.
Can I upload a video instead of a face image?
This page takes a face image plus audio. To change an existing video’s look, use video-to-video.
Does it handle two speakers?
It is built for one face and one voice track — multi-speaker audio gives the mouth ambiguous instructions.
How long can the audio be?
Clips are generated in the 5 to 20 second range shown in the studio; longer speeches need splitting into parts.
Put words in a face
Open LTX-2 Lipsync with an audio file and a portrait.
Keep going
Nearby tools and pages people usually open next.
Other model pages
Each page opens its own model in the same studio.
Pricing

