Wan Animate

Wan Animate takes one character image and a reference performer video, then drives the character with the performer's expressions and movement. You don't rig anything or set keyframes. The model reads the motion from the clip and applies it, which is why you can get a believable result from a plain photo.

Mix Mode replaces a character in existing footage instead of animating a new one. The part that matters here is continuity: the tool preserves the original lighting and color tones, so a swapped character sits in the scene rather than floating on top of it. That keeps the shot usable without a color-grading pass.

The model syncs lip and teeth movement to the audio track, which is what separates a talking character from a puppet. If your performer clip has speech, Wan Animate carries that mouth movement over to your character so the result reads as a natural, lifelike talking performance.

Wan Animate drives body movement with spatially-aligned skeleton signals. Facial expression comes from implicit facial features. In plain terms, it separates "how the body moves" from "how the face looks." That's how it keeps a high degree of control over both instead of blending them into a blur. Why split them at all? Because motion transfer and face cloning fail in different ways, and you want to tune each one on its own.

A relighting layer adjusts the character's appearance to match the environment's lighting and color tone. This is what makes a character look like it belongs in a scene rather than being dropped in, and it's the same idea that does the heavy lifting in Mix Mode swaps.

The whole tool runs on the Wan 2.2 Animate model, which uses a Mixture-of-Experts architecture and high-compression video generation. You don't interact with that directly, but it's the reason the tool can produce high-fidelity character video in a browser without you managing a GPU.