apache/apisix ? reverse-engineered prompt

Build me a self hosted API gateway that I can run locally with Docker, and that can also be used in Kubernetes later.

I want it to route traffic to different upstream services, let me add and change routes without restarting anything, and give me a simple admin API to manage those routes. It should handle load balancing, retries, health checks, rate limiting, authentication, logging, and basic observability. If possible, include support for HTTP, gRPC, WebSocket, TCP and UDP proxying, plus a way to plug in custom behavior for things like request rewriting and response rewriting.

I also want the gateway to work as an AI gateway, so I can send requests to different LLM providers through one place, with fallback options and token based limits. If you need to check current docs or best practices while building it, look them up online.