Personal collection of ComfyUI workflows, experiments, example inputs, and generated previews.
This repository is a toybox for multiple ComfyUI setups, not a single model or a
single workflow family.
Environment note:
Tested on WSL with an RTX 3090. If you are using native Windows, you may need to
change file path separators in load nodes from / to \.
Input image:
Input image
Input audio:
Output video:
MiniMax H3
Reference-to-video pose transfer using a character image and a motion reference
video. The workflow generates a new video with the character following the
reference choreography.