ComfyUI Node
Mpi Wan Masked Video
Wan 2.1 image-to-video conditioning from a whole guide video plus its hole masks (Matrix-3D's masked-video conditioning, ported from ComfyUI-SplatKit, MIT): Wan keeps the known pixels and invents the holes.
Mpi Wan Masked Video
- positive
- negative
- vae
- guide
- holes
- clip_vision_output
- positive
- negative
- latent
◄width1440►
◄height720►
◄length81►
CategoryMpiNodes/Wan
Inputs (9)
| Name | Type | Default | Description |
|---|---|---|---|
| positive | CONDITIONING | — | |
| negative | CONDITIONING | — | |
| vae | VAE | — | |
| guide | IMAGE | The guide video: what is known along the camera path, holes any colour. | |
| holes | IMAGE | One mask a guide frame, WHITE where Wan must invent (a hole), black where the guide is known. | |
| width | INT | 144016–8192 | — |
| height | INT | 72016–8192 | — |
| length | INT | 811–8192 | — |
| clip_vision_outputopt | CLIP_VISION_OUTPUT | — |
Outputs (3)
| Name | Type | Description |
|---|---|---|
| positive | CONDITIONING | — |
| negative | CONDITIONING | — |
| latent | LATENT | — |