theschoolofai/glc_v5 ? reverse-engineered prompt

Build me a local gateway for LLMs and messaging channels that runs on my machine and gives me one place to manage provider keys, route requests to different models, and keep track of usage and costs.

I want a simple web dashboard on localhost with a home page, help page, docs, health check, and a channels setup page where I can paste a control token and save secrets locally without exposing them back to me. It should support a chat endpoint for sending messages to a chosen provider, including multiple Gemini keys that can be used as separate slots, and it should show which actual provider and model handled the request.

Also add budget controls, cost tracking, pricing lookups, semantic caching, and tracing so I can see when a request was blocked, cached, or sent through. It should be able to connect to S16 over a private bridge token and forward channel messages through the same adapter flow. If you need current library docs, look them up online while you build it.