paykaribazaronline/supremeai — reverse-engineered prompt

I want you to turn this repo into a working SupremeAI 2.0 setup that I can run locally and understand without digging through everything myself. The main goal is a usable AI orchestration platform that can route requests across multiple model providers, fall back automatically when one fails, and add strong safety checks so bad or hallucinated responses get filtered or flagged. Please make sure the backend, the web chat, and the admin or studio side all work together in a clean way, with sensible defaults from the example env files.

Please wire up whatever is already in the repo instead of reinventing it, fix broken setup steps, and make the local developer flow simple, ideally one command or a very short sequence. If there are missing pieces, fill them in with practical versions that match the project vision. I’d also like basic visibility into provider status, request routing, failures, and guardrail decisions. If the self improvement or skill system is partially built, get it to a believable working state. Check the project docs and current package docs online if you need to.