orailnoor/cross-platform-llm-client — reverse-engineered prompt
Reverse engineered prompt
Build me a polished cross platform private AI chat app called PrivateLM.
I want users to chat with AI either using local models on their device or cloud providers when they choose. On Android and iPad it should support downloading and running local GGUF models offline, with smart device checks so it recommends safe memory and token settings. On web it can be cloud only.
The app should let people add API keys for OpenAI, Anthropic, Gemini, and Kimi, switch between providers, send text and images in chats, save conversations locally, manage downloaded models, and have a separate task view for more structured AI workflows. Make privacy clear, nothing should leave the device unless cloud mode is selected.
Please make the UI clean and production ready, with settings, model management, chat history, loading progress, errors, and sensible defaults. Include background notifications or task handling where appropriate. Use Flutter for mobile and web, and look up current docs online if you need to.
Want more depth? Deep Reverse