comfyui
https://github.com/comfyanonymous/comfyui
Python
The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported21 Subscribers
Add a CodeTriage badge to comfyui
Help out
- Issues
- QwenImage21Cache device=cpu/auto leaves WDDM shared GPU memory mapped until process restart (Windows, Intel XPU)
- SeedVR2: Abnormal performance fluctuation between versions / drivers
- MiniMax H3 Fun ControlNet aborts natively on GB10 (sm_121) with comfy-kitchen 0.2.35; disabling the CUDA backend resolves it
- Qwen 2.1: Int8_convrot.safetensors model is much slower than Q8 gguf (0.37.0)
- Qwen 2.1 failed to load clip
- Fix native abort: pause block-weight malloc graph while reading Qwen-Image 2.1 prefix cache
- Fix Qwen flash attention CUDA gating
- Qwen Image 2.1 Prompt Expander fails with `deltanet_conv_step launch rejected` on RTX 4090 / Driver 570, but works on L40S / Driver 580
- supports_int8_compute() disables native INT8 on MPS unconditionally — torch._int_mm has been implemented for MPS since Aug 14 2026
- Fix JSONDecodeError when a layer's comfy_quant marker is empty
- Docs
- Python not yet supported