What is this?
ComfyUI is a node-based graph interface for building and running diffusion model workflows: image, video, 3D, and audio generation, editing, and more, all without writing code.
Is this a simple app or more like a visual programming environment?
It's a visual node-graph editor: you wire together nodes (checkpoints, VAEs, samplers, LoRAs, ControlNets) to build custom pipelines, rather than clicking fixed buttons like a typical generator UI.
How do I get it running?
Easiest is the desktop app (Windows/macOS), or a Windows portable build. For manual install on Windows/Linux, git clone the repo, put checkpoints in models/checkpoints and VAEs in models/vae, then install dependencies via pip/terminal.
What do I need before installing it manually?
Python (3.12-3.14, with 3.13 best supported) and PyTorch matched to your GPU vendor: CUDA 13.0+ for recent Nvidia cards, ROCm for AMD, or Intel's XPU build for Arc GPUs.
Any install pitfalls I should watch for?
The Windows portable build isn't recommended for regular users (use the desktop app instead), and if pytorch wasn't built with CUDA you'll hit a 'Torch not compiled with CUDA enabled' error that requires reinstalling torch.
Who is this aimed at?
Visual professionals and power users who want fine-grained control over every model and parameter in generative media pipelines, not just casual prompt-and-click users.
Which platforms and GPUs does it support?
It runs on Windows, Linux, and macOS, with manual install supporting NVIDIA, AMD, Intel, Apple Silicon, and Ascend GPUs.
Does it need an internet connection to run?
The core runs fully offline and downloads nothing unless you request it; you can also disable the optional paid Comfy API nodes to force everything local.
Which models can it run out of the box?
A huge native model library: Stable Diffusion 1.5/SDXL/SD3.5, Flux.1/Flux.2, Qwen Image, Wan video, HunyuanVideo, Hunyuan3D, ACE-Step audio, and many more, plus partner nodes for closed-source models like Nano Banana and Seedance.
Can I save and share my workflows?
Yes, workflows save and load as JSON, and ComfyUI can even recover complete workflows and seeds embedded in generated media files.
Is there a cloud option if I don't have a GPU?
Yes, Comfy Cloud is the official paid cloud version for people who can't afford local hardware.
Can I extend it with my own nodes?
Yes, ComfyUI supports custom nodes to extend functionality, and model storage locations can be configured via extramodelpaths.yaml.
What are similar projects?
krita-ai-diffusion, stable-diffusion-webui, ComfyUI-ALLinONE-MinimaxH3, ai-toolkit, stable-diffusion, DiffSynth-Studio, generative-models, Fooocus
Which topics is it in?
image-to-image-translation, inpainting, 3d-generation, rocm, upscaling, diffusion, comfyui, audio-generation, amd, stable-diffusion, gpl, text-to-image