FastMCP
A Python MCP framework for building local or remote tool servers, connecting clients, and adding interactive interfaces to AI conversations.

FastMCP is a Python framework for developers who want AI agents to use their applications and data through the Model Context Protocol (MCP). It covers both publishing tools and connecting to them, with support for servers that run locally or remotely.
- Tools and data: Expose Python functions as agent tools, share data as resources, and provide reusable prompts.
- Client connections: Call MCP servers from Python, the command line, or another MCP application, with authentication and connection management handled by the framework.
- Interactive results: Return forms, tables, and charts within a conversation alongside a tool's results.
- Input and output checks: Generate schemas from Python functions and validate the data passed between agents and tools.
FastMCP is open source under Apache 2.0 and maintained by Prefect. The same team maintains an official TypeScript counterpart for servers, clients, and interactive apps.
Prefect Horizon is a separate enterprise gateway for team deployment and access management. Its capabilities include a private server registry, SSO, permissions for individual tools, audit logs, and endpoints that group approved tools.