kijai /ComfyUI-WanVideoWrapper

About ComfyUI-WanVideoWrapper

Puts WanVideo and related video models to work inside ComfyUI. ComfyUI is a node-based app for image and video work. The author calls it a personal sandbox, always in progress and prone to issues. New accounts can't post issues right now.

Best for: ComfyUI users who want to try new Wan-family video models before they reach native ComfyUI.

• Adds wrapper nodes for WanVideo and many related video models
• Swaps blocks and offloads work to fit big models into small VRAM
• Keeps unmerged LoRA weights (small add-on models) inside the blocks
• Generates long clips with context windows, frame by frame
• Loads GGUF models through the main model loader
• Works with ComfyUI's own text encoder and clip vision loaders

1. 1
   Clone the repo into your ComfyUI custom_nodes folder.
2. 2
   Install dependencies with the command below.

   pip install -r requirements.txt

   Copy
3. 3
   On a portable install, run the command below.

   python_embeded\python.exe -m pip install -r ComfyUI\custom_nodes\ComfyUI-WanVideoWrapper\requirements.txt

   Copy
4. 4
   Put text encoders in ComfyUI/models/text_encoders, clip vision in ComfyUI/models/clip_vision, the transformer in ComfyUI/models/diffusion_models, and VAE in ComfyUI/models/vae.

• ComfyUI-WanVideoWrapper vs ComfyUI-LTXVideo
• ComfyUI-WanVideoWrapper vs ComfyUI-SeedVR2_VideoUpscaler