Multi-agent (multi-reference) video generation
Generate a fused video from multiple reference images (subject/background) and a prompt; supports @ref_name references.
/openapi/v2/video/fusion/generateBody Parametersapplication/json
1.Basically provide "16:9", "9.16", "4:3", "3:4", "1:1" aspect ratio 2.v6, c1 additionally supports "2:3", "3:2", "21:9"
Video generation duration v4.5 : 5/8 (v3.5 1080p cannot be used 8) v5: 5/8 v5.5/5.6 : 5/8/10 (1080p not available 10)
Support v5.6, v6, c1 / control switch Audio. true: Audio on, false: Audio off
Array of image references (1–3 items) containing "body" or "background" v4.5/v5: Up to 3 pictures v5.5/v5.6/v6: up to 7 pictures
items"v4.5","v5","v5.5","v5.6",,"v6", "c1"
Use @ref_name to describe the scene accurately 1. There must be a space after @ref_name, such as @cat plays 2. The name referenced in prompt must exactly match the ref_name in image_references
"360p","540p","720p","1080p"
Random seed; same seed makes results reproducible
