Apps · Desktop app

Fooocus

by lllyasviel · GPL-3.0 · 53,844 GitHub stars

Type a prompt and get a good image: Midjourney-style simplicity for SDXL, with presets for styles, upscaling and inpainting.

macOSWindowsLinux

Get Fooocus

At a glance

Best for: Beginners who just want good SDXL images.
Inference engine: PyTorch
GPU support: NVIDIA CUDA
AMD ROCm
Apple Metal
CPU only
Opens: Safetensors (original weights)
Features: Chat window
Web chat UI
Platforms: In long-term support: bug fixes only, and SDXL models only.

Sources:

lllyasviel/Fooocus on GitHub

Install

macOS

Download the package from the GitHub README

Windows

Download the package from the GitHub README

Linux

Download the package from the GitHub README

9 models on this site run in Fooocus

Model	Made by	Type	Size	Opens it as	Quality tier
Juggernaut XL v9	RunDiffusion	Image	2.6B	Built-in support	62
RealVisXL V5.0	SG161222	Image	2.6B	Built-in support	62
Animagine XL 4.0	Cagliostro Lab	Image	2.6B	Built-in support	60
Illustrious XL v0.1	OnomaAI	Image	2.6B	Built-in support	60
DreamShaper XL v2 Turbo	Lykon	Image	2.6B	Built-in support	58
Pony Diffusion V6 XL	PurpleSmartAI	Image	2.6B	Built-in support	58
Animagine XL 3.1	Cagliostro Lab	Image	2.6B	Built-in support	56
Stable Diffusion XL 1.0	Stability AI	Image	2.6B	Built-in support	55
SDXL-Lightning (4-step)	ByteDance	Image	2.6B	Built-in support	54

Alternatives

ComfyUI

Draw Things Invoke

SwarmUI Unsloth Desktop SD.Next Forge Neo

DiffusionBee

Krita AI Diffusion

Stability Matrix JoyFusion

Which app should I use?

Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.