golf-mcp/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

What it solves

Golf simplifies the process of building Model Context Protocol (MCP) servers. It removes the need for repetitive boilerplate code by allowing developers to define server capabilities—such as tools, prompts, and resources—as simple Python files organized in a directory structure, which the framework then automatically compiles into a runnable server.

How it works

Golf uses a convention-over-configuration approach. Developers create Python files in specific directories (tools/, resources/, and prompts/), and Golf automatically discovers and parses these files. It uses module docstrings for descriptions and Pydantic models for output schemas. The framework targets FastMCP 4.0.0 and the MCP 2026-07-28 protocol, supporting transports like streamable-http and stdio.

Who it’s for

It is designed for developers building AI agents and MCP servers who want to accelerate development and reduce infrastructure overhead.

Highlights

  • Automatic Discovery: Automatically converts Python files in a directory structure into MCP server components.
  • Enterprise Authentication: Built-in support for JWT, OAuth Server, and development tokens.
  • Telemetry Integration: Native support for OpenTelemetry tracing to monitor server performance.
  • CLI Tooling: Includes a CLI for project initialization (golf init), building, and running servers.

Related

  • Project
    NVIDIA-NeMo/ProRL-Agent-Server

    A reinforcement learning rollout framework that provides rollout-as-a-service for training multi-turn LLM agents across diverse agent harnesses.

  • Project
    brucefan1983/GPUMD

    A high-performance GPU-accelerated molecular dynamics package that supports empirical and neuroevolution machine-learned potentials for materials simulation.

  • Project
    kruzovic7/ai-data-extractor

    A toolkit to extract and normalize local chat history from multiple AI coding assistants into JSONL format for backup, analytics, or fine-tuning.

  • Project
    AgentWorkforce/relay

    Agent Relay is an open‑source toolkit that lets teams run LLM‑based coding agents together with existing tools (GitHub, Slack, Notion, etc.). It provides messaging, virtual‑filesystem integrations, searchable session logs, and TypeScript‑defined deterministic flows, enabling auditable, rule‑driven automation of code‑generation and review tasks.