Wan Animate Move
Use the driving video as a motion reference for the character in your image. It works well for dance clips, gestures, talking-head movement, and stylized character performances.
Transfer motion or replace a character from one image and one driving video
Turn a portrait or character image into a performance guided by your reference video. Choose Move for motion transfer or Replace to swap the person while keeping the source action and timing.
Choose your goal
Example: make the character in your image follow a dance or gesture from the video.
Keep it short and specific. Maximum 500 characters.
Output resolution
Create with Wan 2.2 Animate Move or Replace. The quote uses the selected resolution and verified MP4 duration, with a 3-second billing minimum.
Source: Wan Animate project demo
Choose the result you need
Each workflow uses a video and an image, but changes a different part of the result. Start with face swap if keeping the original body movement is your priority.
Planning a photo-to-performance clip? Follow the Raindance AI guide to prepare your character image and reference motion.
Use the driving video as a motion reference for the character in your image. It works well for dance clips, gestures, talking-head movement, and stylized character performances.
Use the image as the new character reference inside the driving video. Choose it when the source timing and scene work, but the on-screen person needs to change.
Change only the face while retaining the original body, clothing and scene. Clear, visible faces are important; exact identity matching is not guaranteed.
Edit the performer or appearance using a prompt and reference image. Suitable for stylized edits; generated movement may differ from the source. Requires a 3–10-second video.
Trending now
The newest Wan animation discussions center on end-to-end character animation, identity preservation, text-driven viewpoint control, and real-time character animation. Wan2.2-Animate-2-14B, Hugging Face, ComfyUI workflows, tutorials, and local deployment now appear together across current searches and discussions.
For creators, the practical idea stays simple: provide a strong character reference, use a clean driving video, and judge motion, identity, framing, and temporal consistency as separate quality signals.
View the official Wan Animate 2 model cardImage-to-animation workflow
Build a clean, full-body character reference in the Wan 3 image generator first. Then bring that image here with a driving video to create a repeatable character-animation workflow without setting up a local ComfyUI graph.
Create a character with Wan 3 ImageChoose a clear MP4 up to 30 seconds and 10MB. Keep the performer visible and avoid abrupt cuts for more consistent motion.
Use a sharp PNG, JPG, or WebP with a readable face, body shape, and outfit. Match the source framing when possible.
Choose a mode, follow its input limits and review the credit quote. Resolution is selectable only for Wan Animate; video editing also requires a short prompt.
Move and Replace use Wan 2.2 Animate. Face swap uses a dedicated face-swap model. Video character editing uses Kling O3 Standard. Each mode shows its own input limits and credit quote.
A driving video is the motion reference. Its body movement, timing, gestures, and camera framing guide the generated character animation.
Move transfers motion from the driving video to the character image. Replace uses the character image to replace the person in the source video.
All modes accept one MP4 and one PNG, JPG or WebP, each up to 10MB. Wan Animate and face swap allow up to 30 seconds. Video editing requires 3–10 seconds with both dimensions 700–3840px. Only Wan Animate offers 480p or 720p; other modes set output size automatically.
Credits depend on the selected mode and verified video duration, with a 3-second minimum. Wan Animate also considers the chosen resolution. The quote appears before submission. Failed jobs return the charged credits; visual quality is evaluated separately.
No. This is a hosted browser workflow, so you can upload inputs and run the animation without installing a local ComfyUI workflow.
Upload a clean reference image and a short driving video, then choose the animation mode that matches your shot.
Start animating