aDeLaLaHmaDi/freellmapi ? reverse-engineered prompt

Build me a local app that gives me one OpenAI compatible endpoint, but routes requests across a bunch of free LLM providers behind the scenes. I want to be able to add my own API keys, have them stored securely, and then use the same server with OpenAI style clients for chat, models, responses, embeddings, images, and audio. If one provider is rate limited or fails, it should automatically try another one so requests keep working.

Please include a simple web dashboard where I can manage keys, see the available models, and turn providers on or off. It would be great if the app could also support Claude style and other compatible clients, plus a desktop app and Docker setup so it is easy to run locally. If you need to look up current API docs online to make the integrations accurate, do that. Keep it focused on personal experimentation and make the default experience easy to start with.