Node-based workflows for generative AI images, video and audio
ComfyUI provides a browser interface for building and running generative AI
workflows with local models. Models are supplied separately.
NVIDIA GPU support is provided by the torch-cuda component.
The server runs at http://127.0.0.1:8188. Startup, listening address
and port can be configured through snap settings. Models,
inputs, outputs and settings are stored in /var/snap/comfyui/common.
Install comfyui-desktop for the desktop application.
The comfyui command remains available for a separate foreground server.