nomic-ai/gpt4all — reverse-engineered prompt

Build me a polished local AI chat app like GPT4All that lets people run large language models privately on their own computer with no cloud dependency and no API key. I want a desktop app for Windows, macOS, and Linux where someone can download a model, start chatting right away, and keep everything on device. Include a clean chat interface, a model browser or downloader, support for modern GGUF style models, and a way to chat with personal documents locally like a private docs assistant.

Also include a simple Python package or example so developers can load a model and generate text from code, plus support for chat sessions. If it fits the project, include optional local acceleration for compatible hardware, but it should still work on normal laptops without a GPU.

Please make it feel production ready, easy to install, and commercially usable as an open source project. Look up the current docs online if you need to.