WAN 2.2-S2V
Turns speech recordings into talking avatar videos with lip sync.
Introduction
WAN 2.2-S2V is a speech-to-video tool that turns an audio recording into an avatar video with lip sync and natural facial movement. You can pick a realistic avatar or upload your own photo. The underlying model is released under the Apache 2.0 licence and is available on Hugging Face and ModelScope, and the site says it can produce 720p video in under ten minutes.
- Added on
- Nov 2, 2026
- Website
- wan-s2v.com
Product information
What is WAN 2.2-S2V?
A speech-to-video tool that creates avatar videos from audio, backed by an open-source model.
How to use WAN 2.2-S2V?
Record or upload speech audio, choose an avatar style or upload a photo, let the model generate synchronised video, then download the result.
Core features of WAN 2.2-S2V
- Speech-driven avatar videos with lip sync
- 27B-parameter Mixture-of-Experts model
- 720p output in under ten minutes
- Support for more than 40 languages
- Apache 2.0 licensed model
Use cases of WAN 2.2-S2V
- 1Lectures and tutorials
- 2Presentations and corporate training
- 3Marketing and product videos
- 4Podcast visuals
Frequently asked questions about WAN 2.2-S2V
What speech formats and languages are supported?+
The site lists MP3, WAV, M4A and FLAC audio and says 40+ languages are supported.
Can I customise avatars with my own photos?+
Yes, you can upload a photo to create a personalised avatar.
Which category is WAN 2.2-S2V in?+
WAN 2.2-S2V is listed under AI Avatar Video.