Stable Diffusion web UI by AUTOMATIC1111 is an open-source web interface for running Stable Diffusion image generation on your own computer. Built with the Gradio library, it provides text-to-image and image-to-image modes, inpainting, outpainting, upscaling and a settings page for fine control. It is free to download from GitHub, with a one-click install and run script, though Python and git must be installed first. Support extends to Windows, Linux and Apple Silicon.
Hobbyist artists, tinkerers, researchers and developers who prefer local generation over hosted services are the main audience. Because it runs offline, users keep control of models, prompts and outputs, and extensions let them add features that the community builds. Users with modest hardware can still work, since the project advertises support for 4GB video cards, with reports of 2GB working. Those who want fine control over parameters, seeds and samplers will find more switches here than in simpler apps.
Typical work includes iterating on prompts with attention weighting, fixing faces with GFPGAN or CodeFormer, and upscaling results with RealESRGAN, ESRGAN, SwinIR or LDSR. Artists use inpainting to repair parts of an image and outpainting to extend a canvas, while prompt matrix and X/Y/Z plots compare how settings change results. Textual inversion lets users train embeddings on 8GB of VRAM, and generation parameters are saved inside the PNG or EXIF data so any image can be reproduced later by dragging it into the PNG info tab. Tiling support creates tileable textures, and a live preview shows progress during generation.
