Model Context Protocol (MCP) Experimental Garden - The most developer-friendly experimental playground for MCP implementations
MCP-Garden is the most developer-friendly experimental garden for Model Context Protocol implementations. We provide runnable, comparable, and reusable experimental projects for servers, clients, and LLM integrations, lowering the barrier for developers to get started and innovate.
Model Context Protocol (MCP) is an open standard for connecting AI assistants with data sources and tools. It enables secure access to external data sources, execution of operations, and integration with various services.
A tool that automatically generates Python packages from mcp.json configuration files. Each MCP server corresponds to an independent Python package, and each tool corresponds to an async function.
Key Features:
- Generate independent Python packages for each MCP server
- Lazy connections with global connection pool
- Auto-generated type hints from JSON schemas
- Support for stdio and HTTP transports
A streamlined SDK for converting FastAPI applications to MCP servers using FastMCP's experimental OpenAPI parser.
Key Features:
- Convert FastAPI endpoints to MCP tools automatically
- Support both separate and integrated server modes
- MCP Official Docs
- MCP Specification
- Awesome MCP - Curated list of MCP projects, tools, and resources
MIT License - see LICENSE for details.