Back to Nodes
Wan 2.2 Animate Replace

Wan 2.2 Animate Replace

Official

Replaces a character in a video scene with a new character image. Part of the advanced Wan2.2 video model family. Ideal for creative video editing and

Nodespell AI
AI / Video / Alibaba

Replaces a character in a video scene with a new character image. Part of the advanced Wan2.2 video model family. Ideal for creative video editing and content creation.

Model Overview

A video model that replaces a character in a scene with a new character image, using the Wan2.2-Animate technology.

Best At

Replacing a character in a video with a new image while preserving the video's motion, expressions, and (optionally) audio. Works well when the reference frames show the character clearly.

Limitations / Not Good At

Performance may degrade if the reference frames are low-quality or if the character is small or partially obscured in the input video. The model requires at least 1 or 5 reference frames (set by refert_num). Complex camera movements might affect the alignment.

Ideal Use Cases

  • Replacing actors in existing footage without reshooting.
  • Testing different character looks in a video.
  • Creating custom avatars that mimic the movements of a real person.

Input & Output Format

Input: Video (URI) and Character Image (URI) for replacement. Optional parameters include resolution, reference frame count, and frame rate.
Output: Generated video (URI).

Performance Notes

This model can be run in "go_fast" mode for faster generation. The base Wan2.2 model family is designed to run efficiently on consumer-grade GPUs (like 4090) for 720P video generation.

Inputs (2)

Video

String

Input video containing the character to be replaced

Min: 0Max: 100

Character Image

String

New character image to replace the original character

Min: 0Max: 100
Parameters (5)

Seed

Number

Random seed. Leave blank for random

Default: -1

Go Fast

Boolean

Go fast

Default: true

Resolution

String

Resolution for processing

Default: 720

Merge Audio

Boolean

Merge audio from input video into output (experimental - may not sync perfectly)

Default: true

Frames Per Second

Number

Frames per second for output video

Default: 24
Outputs (1)

Output

Inferred

Output

Nodespell

Nodespell

📍 London

Building the future. Join us!

Type

Node

Status

Official

Package

Nodespell AI

Category

AI / Video / Alibaba

Input

VideoImage

Output

Video

Keywords

Image EditVideo Edit
Use in Workflow