0xzr/freellmpool ? reverse-engineered prompt

Build me a free LLM gateway that can sit in front of a bunch of chat models and providers, then give me one simple local endpoint I can point my apps at. I want it to work from the command line, as a Python library, and as a local proxy, with an easy first run setup that checks what keys or tools I already have and tells me what to do next.

It should automatically use only the routes that are actually available, fall back to another model if one is rate limited or down, and start without any API key when a free keyless option is available. I also want a way to discover and import local models like Ollama or LM Studio without scanning my whole network, plus a couple of ready made profiles for coding tools like Cursor, Claude Code, and OpenCode. If you need to check current provider docs online, do that too.