AUTOMATIC1111/stable-diffusion-webui ? reverse-engineered prompt

I want to build a comprehensive web UI for Stable Diffusion that I can run on my local machine. The main purpose is generating images from text prompts, so it needs a primary text-to-image tab. It should also have an image-to-image feature where I can upload a picture and have the AI modify it based on my text.

Let's make it powerful. The interface should have both a positive and a negative prompt field. I want advanced editing tools like inpainting, where I can mask part of an image to regenerate just that area, and outpainting to expand the canvas. Please also add an "extras" section with utilities for fixing faces and upscaling images to a higher resolution.

Beyond the basics, it should support loading different models and checkpoints, and even let me merge a few models together. A really cool feature would be a "CLIP interrogator" that can look at an image and guess the prompt that created it. For the tech, let's use Python and the Gradio library to build the interface.