MrFadiAi/free-llm-gateway ? reverse-engineered prompt

Build me a local API gateway for AI models that feels like one simple OpenAI style endpoint, so I can point my apps at one base URL and use one master key. It should let me drop in free provider keys from a config file, then automatically route requests to whichever model works best, and fall back to another provider if one is slow, rate limited, or fails. I also want a web dashboard where I can see available models, test prompts, watch usage and rate limits, and manage keys and fallback order without restarting anything. It should support streaming responses, basic analytics, request logs, and a simple playground to try models from the browser. If you need to, look up current docs online for any provider details, but keep the experience simple and make it easy to run locally or with Docker.