bentoml/OpenLLM ? reverse-engineered prompt

Build me a Python tool that lets me run open source language models locally or in the cloud with one simple command. I want to be able to start a model server, talk to it through an OpenAI compatible API, and also open a simple chat page in the browser. It should support common models like Llama, Mistral, Qwen, Phi, and DeepSeek, and it should also let me plug in my own custom models from a repository if I want to.

Please make the experience feel easy for someone who just wants to say serve or run and get a working model endpoint. I’d like model listing, model info, and a basic hello or getting started command too. If you need to check current docs or model names online, go ahead and do that. Keep it practical, polished, and ready to use with common tools that already understand OpenAI style APIs.