spacepirate15/quantum-free-router — reverse-engineered prompt

Build me a local router for free LLM APIs that I can point coding or research agents at with one OpenAI compatible base URL, like http://127.0.0.1:4000/v1. I want it to run on Linux or WSL2, install in one command, keep my real API keys in a local user config file, and run as a background service I can restart and health check easily.

The main goal is reliability. If one free provider is down, slow, out of quota, or timing out, the router should automatically fall through to other working free models instead of killing the whole task. Please include a curated provider and model catalog, sensible timeout and retry behavior, simple health and certification scripts, and clear docs for setup, configuration, limits, troubleshooting, and provider notes. Add a couple of examples, like curl and a small Python client, so I can test it fast.

Please keep the default path tuned for long running agent work, and look up current docs online if you need to.