ComfyUI Node
Continue From Last Clip
On clip 1 this does nothing. On every later clip it copies the saved end of the clip before into the start of this latent and locks it, so the video carries on without a jump.
Continue From Last Clip
- chain
- latent
- first_image
- vae
- latent
- start_image
◄namelow res►
◄keep_look_of_clip_1true►
CategoryTrent/MiniMax
Inputs (6)
| Name | Type | Default | Description |
|---|---|---|---|
| chain | H3_CHAIN | From Video Chain. | |
| latent | LATENT | This clip's latent before sampling: an empty H3 AV latent, or the upscaled latent in a second pass. | |
| name | STRING | low res | Same name as on Remember Clip Ending. Use one name per pass size. |
| keep_look_of_clip_1 | BOOLEAN | true | Copies the first frame of clip 1 into the hidden first frame of every clip, so colours and faces drift less over a long chain. The frame is cut away again, so it never shows. |
| first_imageopt | IMAGE | Start image of clip 1. | |
| vaeopt | VAE | Video VAE. Needed only for start_image on later clips. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| latent | LATENT | Latent for the sampler. Clip 1: unchanged. Later clips: starts with the end of the clip before. |
| start_image | IMAGE | first_image on clip 1, the first copied frame on later clips. Connect to first_frame of MiniMax H3 Image to Video. |