Fooocus is free, open-source image generation software that runs offline on your own computer. It is built on the Stable Diffusion XL architecture and rethinks the interface so users concentrate on prompts and images rather than parameters. Like online generators such as Midjourney, it avoids heavy prompt engineering, helped by an offline GPT-2 based prompt processing engine. The project is non-commercial, needs no subscription, and installs in fewer than three clicks between download and the first image.
The software suits hobbyists, digital artists and developers who want local control over image generation without paying for credits or uploading prompts to a service. Minimum GPU memory is 4GB on Nvidia hardware, so many consumer graphics cards can run it. People comfortable with Python can install it on Linux, Windows or Mac through Anaconda, a virtual environment or native Python, and Docker and Colab options exist for those who prefer containers or cloud notebooks. AMD GPU instructions are available for both Linux and Windows.
Typical uses include generating concept art from a short description, exploring styles with the bundled presets and style library, and running batches with wildcards. Users can protect a shared instance with authentication and customize the interface. The project is now in limited long-term support, with only bug fixes planned and no current migration to newer architectures, so people who want models such as Flux are pointed to alternatives like WebUI Forge, ComfyUI or SwarmUI. Maintainers also warn that many fake Fooocus websites exist, and the GitHub repository is the only official source.
