golf
Production-Ready MCP Server Framework • Build, deploy & scale secure AI agent infrastructure • Includes Auth, Observability, Debugger, Telemetry & Runtime • Run
What is golf?
Production-Ready MCP Server Framework • Build, deploy & scale secure AI agent infrastructure • Includes Auth, Observability, Debugger, Telemetry & Runtime • Run real-world MCPs powering AI Agents
Overview of golf
What it is
Golf is a framework for building MCP server applications. It allows developers to define server capabilities—tools, prompts, and resources—as Python files within a conventional directory structure. Golf automatically discovers, parses, and compiles these components into a runnable MCP server. It targets FastMCP 4.0.0 and the MCP 2026-07-28 protocol. Version 0.2.0 includes enterprise-grade authentication (JWT, OAuth Server, development tokens), built-in utilities for LLM interactions, and automatic telemetry integration.
Key features
- Automatic discovery and compilation of tools, prompts, and resources from Python files in a conventional directory structure.
- Enterprise-grade authentication: JWT, OAuth Server, and static development tokens.
- Built-in utilities for LLM interactions: elicit, sample, get_current_context.
- Automatic OpenTelemetry tracing integration.
- Anonymous CLI telemetry to understand framework usage (can be disabled).
- Configuration via golf.json for server name, host, port, transport, and tracing settings.
Use cases
- Building production-ready MCP servers for AI agents.
- Rapid prototyping of tools, prompts, and resources with minimal boilerplate.
- Deploying secure AI agent infrastructure with authentication and observability.
- Integrating with LLM workflows that require elicitation and sampling.
Getting started
bashpip install golf-mcpgolf init your-project-namecd your-project-namegolf build devgolf run
License and pricing
This project is licensed under Apache-2.0 and is open source.
Based on the project's public README and repository metadata.
