PrefectHQ/fastmcp ? reverse-engineered prompt

I want to build a Python framework that makes it super easy for developers to connect their code and data to LLMs. Let's call it FastMCP.

The main idea is that a developer can just take a normal Python function, add a simple decorator to it, and the framework automatically turns it into a tool that an LLM can use. It should handle all the hard parts in the background—like generating schemas, validating inputs, and managing the server communication based on the Model Context Protocol. The developer shouldn't have to worry about the protocol details; they should just focus on writing their tool's logic.

The framework needs to support building the tool servers, creating clients to connect to them, and also a way to build simple interactive UIs for the tools that can appear right inside a chat conversation. The goal is to make this the standard, go-to library for this kind of work in Python, where best practices are built-in.