izaart95-jpg/GLM-Free-API ? reverse-engineered prompt

Build me a Go service that acts as a drop in proxy for chat.z.ai so I can use GLM models from tools that already speak OpenAI or Anthropic APIs.

I want one server with /v1/chat/completions and /v1/messages, plus a model list and a simple health check. It should handle both streaming and normal replies, use a bearer token for auth, and work without any browser automation. If possible, make it support agent mode too, so tool calls can be translated cleanly and sent back in a format the client understands.

It should also keep track of available models from Z.AI, fall back to a small built in list if needed, and support basic config through env vars like port, host, timeout, and optional ZAI token. I’d like a clean startup, helpful status info, and a simple way to run it locally with no complicated setup. If you need to check current docs or API behavior, look them up online first.