ComfyUI Node
Motion Bridge Save ????
A ComfyUI node in IAMCCS/Wan with 6 inputs and 0 outputs.
Motion Bridge Save ????
- latent
- images
◄namesegment_bridge►
◄tail_frames2►
◄save_preview_pngtrue►
◄flush_vramtrue►
CategoryIAMCCS/Wan
Inputs (6)
| Name | Type | Default | Description |
|---|---|---|---|
| latent | LATENT | — | |
| images | IMAGE | — | |
| name | STRING | segment_bridge | Bridge file name (no extension). Use descriptive names like 'scene01_seg2_to_seg3'. Saved to output/motion_bridges/<name>.safetensors |
| tail_frames | INT | 21–16 | How many latent frames to slice from the tail of 'latent'. Must match 'motion_latent_count' on WanImageMotion* in the next workflow. |
| save_preview_png | BOOLEAN | true | Also save the last frame as <name>_lastframe.png alongside the .safetensors. Useful for visual inspection and as a fallback LoadImage source. |
| flush_vram | BOOLEAN | true | After saving, move input tensors to CPU and call ComfyUI soft_empty_cache + torch.cuda.empty_cache(). Frees CUDA reserved blocks before the next KSampler runs. Disable only if you need the tensors to stay on GPU for another branch that runs immediately after this node. |
Outputs (0)
No outputs