wan2.7-r2v
Wan2.7-R2V is a reference raw video model that supports three modal inputs: text, image, and video. It can reproduce the appearance, timbre, and movements of characters from reference materials to generate natural and realistic performance videos. Its core capability lies in maintaining character consistency: it can provide up to 5 reference materials, support single or multi-person performances, and the reference video can also reproduce voice timbres. The model supports multi-storyboard narrative. Users can describe the shot logic through time periods, and the model automatically identifies the storyboard structure. In addition, the sound can be set independently through the reference_voice audio file. If not specified, the audio in the reference video will be used by default. The output supports 720P/1080P resolution, the screen ratio can be 16:9, 9:16, 1:1, etc., and the video duration is up to 10 seconds.
- Context
- 0 tokens
- Max output
- 0 tokens
- Input
- llm image video
- Output
- video
Pricing
Price: $0.15 / request
720P reference_to_video : $0.09999999999999999 /
720P reference_to_video : $0.09999999999999999 /
1080P reference_to_video : $0.15 /
1080P reference_to_video : $0.15 /
Endpoints
- openai