Actions12
Video → Replace Person in Video
AI-generatedSummary
Replaces a person in a video with a character from a reference image using the Wan2.1 model.
Inputs
- videoBinaryProperty (required) — Name of the binary field containing the input video file (MP4, MPEG, QuickTime, AVI, WMV, OGG; max 8s).
- refImageBinaryProperty (required) — Name of the binary field containing the reference character image (JPG, JPEG, PNG, GIF, BMP, WebP; max 10MB).
- prompt — Optional text description to guide the replacement (default empty).
- ratio (required) — Aspect ratio of the output video (original, square, landscape, portrait).
- options (required) — Advanced options including resolution presets, random seed (-1 for random), and wait timeout.
Output shape
a single record with the original input data paired to the output.
The output retains the input item structure (pairedItem). The process uses a webhook to handle the async generation.
Examples
Example 1: Replace a person in a 5-second landscape video with a specific character.
Input: video field 'data', ref image field 'data1', ratio 'landscape', seed -1, timeout 120s.
Example 2: Generate a square video replacement using a custom resolution.
Input: video field 'data', ref image field 'data1', ratio 'square', resolution 512x512, seed 12345.