# ComfyUI ComfyUI is the most powerful and modular diffusion model GUI, API, and backend with a graph/nodes interface. It lets you design and execute advanced [[Stable Diffusion]] pipelines using a visual flowchart-based interface instead of code. Available on Windows, Linux, and macOS. ## Philosophy Unlike traditional linear UIs, ComfyUI gives you full control—branch, remix, and adjust every part of your workflow at any time. It's 100% open source and designed to run locally for faster iteration, lower costs, and complete control. ## Key Features ### Visual Workflow Editor - **Nodes/graph interface**: Build AI workflows visually by connecting nodes on a canvas - **Reusable workflows**: Save, share, and reuse entire workflows; exported files carry metadata for instant rebuild - **Live preview**: See results in real time as you adjust workflows - **Smart execution**: Only re-executes parts of the workflow that change ### Supported Models **Image Models:** - SD 1.x, SD 2.x, SDXL, SDXL Turbo - Stable Cascade, SD3, SD3.5 - Pixart Alpha/Sigma, AuraFlow, HunyuanDiT - [[Flux]], Flux 2, Lumina Image 2.0 - HiDream, Qwen Image, Hunyuan Image 2.1 **Video Models:** - Stable Video Diffusion, Mochi, LTX-Video - Hunyuan Video, Wan 2.1/2.2 **Audio Models:** - Stable Audio, ACE Step **3D Models:** - Hunyuan3D 2.0 ### Technical Features - **Smart memory management**: Run large models on GPUs with as low as 1GB VRAM - **CPU fallback**: Works without GPU (slow but functional) - **Model support**: ckpt, safetensors, standalone diffusion models, VAEs, CLIP - **LoRA, LoHa, LoCon**: Full support for model fine-tuning adapters - **ControlNet and T2I-Adapter**: Precise control over generation - **Inpainting, outpainting, upscaling**: Complete image editing toolkit - **Embeddings/Textual inversion**: Custom concept injection - **Workflow metadata**: Load full workflows from generated PNG, WebP, FLAC files ### Extensibility - **Custom nodes**: Extend ComfyUI by building your own nodes - **API nodes**: Optional integration with paid external providers - **Config file**: Set custom search paths for models ## Installation Options 1. **Desktop Application**: Easiest way (Windows & macOS) 2. **Windows Portable Package**: Latest commits, completely portable 3. **Manual Install**: Supports all OS and GPU types (NVIDIA, AMD, Intel, Apple Silicon, Ascend) ## Release Cycle Weekly releases targeting Monday, with three interconnected repositories: - ComfyUI Core (stable versions) - ComfyUI Desktop (desktop builds) - ComfyUI Frontend (UI updates) License: GPL-3.0 ## References - Official Website: https://www.comfy.org/ - Documentation: https://docs.comfy.org/ - Source code: https://github.com/Comfy-Org/ComfyUI - Examples: https://comfyanonymous.github.io/ComfyUI_examples/ - Download: https://www.comfy.org/download