invoke-ai/InvokeAI ? reverse-engineered prompt

Build me a local AI image app that feels polished and easy to use for artists and hobbyists.

I want a web based interface where I can type prompts and generate images, then quickly refine them, edit parts of an image with inpainting and outpainting, and keep iterating without losing my work. Please include a canvas for drawing and editing, a workflow area for building node based generation pipelines, and a gallery or board system so I can organize, revisit, and remix images later. It should support common Stable Diffusion style models and other modern image models, with a model manager, upscaling tools, and basic metadata so I can see what settings were used. Make it run locally, feel responsive, and work well on desktop browsers.

If you need to look up current model or install details online, go ahead and do that.