Wan 2.2 Animate Move model transfers motion and expressions from a reference video to a character in a static image, bringing the image to life with dynamic performance. Ideal for dance replication, complex body movement transfer, and low-cost motion capture alternatives. Processing time: several minutes.
API Key authentication. Format: Bearer YOUR_API_KEY.
Input character image URL (HTTP/HTTPS, publicly accessible, no Chinese or non-ASCII characters). Format: JPG/JPEG/PNG/BMP/WEBP. Size: width and height both in [200, 4096] pixels, aspect ratio 1:3 to 3:1. File size: ≤ 5MB. Content requirement: single person facing camera, full face visible without obstruction, moderate proportion in frame (avoid too large or too small)
"https://example.com/person.jpg"
Reference motion video URL (HTTP/HTTPS, publicly accessible, no Chinese or non-ASCII characters). Format: MP4/AVI/MOV. Duration: 2-30 seconds. Size: width and height both in [200, 2048] pixels, aspect ratio 1:3 to 3:1. File size: ≤ 200MB. Content requirement: single person facing camera, full face visible without obstruction, moderate proportion in frame. Tip: Higher resolution and frame rate of reference video can improve output quality
"https://example.com/dance.mp4"
Model service mode. Options: (1) 'wan-std': Standard mode - faster generation speed, cost-effective, suitable for quick preview and basic animation scenarios; (2) 'wan-pro': Professional mode - higher animation smoothness and better quality, longer processing time and higher cost, suitable for final production output
wan-std, wan-pro "wan-std"
Whether to perform image detection. When true (default), the API validates the input image for quality and content requirements. When false, skips image detection for faster processing. Set to false only when you're confident the image meets all requirements
true