Portrait image + audio → talking video, 480p or 720p.
Model Category
Lip Sync Models — 15 Ways to Match Audio to a Face
9 named models across two input modes — animate a portrait from audio, or re-sync the lips on an existing video. 15 models in total, no content filters.
15 Lip Sync Models9 Named Across Two Input Modes
Lip sync models drive a talking video from an audio track, either animating a still portrait or re-syncing the lips on an existing clip. The category ships 15 models in total, 9 of them named with dedicated endpoints split across image-based and video-based modes. This page covers the models themselves — see the Lip Sync Studio page for the generation workflow.
Lip Sync Models
15 Lip Sync Models, One Open Generative AI Catalog
Every model below is listed by the input mode it takes — a portrait image plus audio, or an existing video plus audio. The Lip Sync Studio page covers how a generation run is put together; this page stays on the models.
Portrait image + audio → talking video, 480p or 720p.
Portrait image + audio → talking video, up to 1080p.
Portrait image + audio → talking video, up to 1080p.
Re-sync the lips on an existing video with new audio.
Video + audio → lipsync video.
Video + audio → lipsync video.
Video + audio → lipsync video.
Video + audio → lipsync video, 480p or 720p.
Hosted vs. run it locally
Hosted Today, With No Local Lip Sync Path
Open Generative AI documents local inference for image and video generation only, so these models run through the hosted API — the same one the rest of the catalog uses.
Hosted
All 15 lip sync models — 9 of them named, across image-based and video-based modes — run through the hosted API with no local setup.
Local Inference
Local inference (sd.cpp and Wan2GP) is documented for image and video generation only; the source project lists no local path for lip sync, so it runs hosted today.
Compare
Three Lip Sync Models, Side by Side
A quick look at three representative models in the catalog — one per input mode, plus the highest-resolution option.
| Model | Input mode | Resolutions |
|---|---|---|
| Infinite Talk | Portrait image + audio | 480p, 720p |
| LTX 2.3 Lipsync | Portrait image + audio | 480p, 720p, 1080p |
| Sync Lipsync | Existing video + audio | Not specified |
Swipe to see the full table
FAQ
Lip Sync Model FAQ
15 in total; 9 of them are named with dedicated endpoints across two input modes — 4 image-based models and 5 video-based models.
Explore more
More Open Generative AI Model Categories
Lip sync pairs closely with the categories below. For the generation workflow itself — choosing an input mode, preparing the audio track, running the job — read the Lip Sync Studio page instead.
Lip Sync a Video With Open Generative AI — Free, Right Now
15 lip sync models, no content filters, no subscription. No account required to try it.
Try Free