Wan 2.6 Reference-to-Video model generates videos from reference URLs (images/videos) with multi-character interaction and role-playing capabilities. Generates silent videos by default.
API Key authentication. Format: Bearer YOUR_API_KEY.
Video content description (supports Chinese and English, up to 1500 characters). Can reference characters using 'Character1/Character2' format
1 - 1500"Character1 drinking bubble tea while dancing to music"
Reference URLs array (images + videos ≤ 5)
1 - 5 elementsReference media URL (HTTP/HTTPS/OSS)
["https://example.com/character.jpg"]Negative prompt describing unwanted content
500"blurry, distorted, low quality"
Video resolution (format: width*height)
1280*720, 720*1280, 960*960, 1088*832, 832*1088, 1920*1080, 1080*1920, 1440*1440, 1632*1248, 1248*1632 "1920*1080"
Video duration in seconds
2, 3, 4, 5, 6, 7, 8, 9, 10 5
Shot type (single-shot or multi-shot narrative)
single, multi "single"
Random seed for reproducibility
0 <= x <= 214748364742