Redbook Search Comment MCP

🚀 MCP Python SDK Documentation

The MCP Python SDK provides a convenient way to integrate your application with large language models (LLMs) using the Model Context Protocol (MCP). This documentation will guide you through the installation, usage, and key concepts of the MCP protocol.

🚀 Quick Start

MCP (Model Context Protocol) is a protocol designed to connect with large language models (LLMs). Through MCP, you can integrate your application with LLMs to provide users with a more powerful interactive experience.

Example: Creating a Basic MCP Server

Here is a simple example of an MCP server:

from mcp import ServerSession, StdioServerParameters, types
from mcp.server.stdio import stdio_server

async def run():
    async with stdio_server() as (read_stream, write_stream):
        async with ServerSession(
            read_stream,
            write_stream,
            initialization_options=types.InitializationOptions(
                server_name="my-server",
                server_version="1.0.0",
                capabilities=server.get_capabilities(),
            ),
        ) as session:
            await session.run()

if __name__ == "__main__":
    import asyncio

    asyncio.run(run())

📦 Installation

Installation

To install the MCP Python SDK, run the following command:

pip install mcp-sdk

📚 Documentation

Core Concepts of the MCP Protocol

MCP defines three core concepts:

MCP Server Capabilities

The MCP server declares its capabilities during initialization:

Example: Developing an MCP Server

Example: Creating a Server that Provides File Resources

from mcp import ServerSession, StdioServerParameters, types
from mcp.server.stdio import stdio_server

async def handle_list_prompts() -> list[types.Prompt]:
    return []

async def handle_get_prompt(name: str, arguments: dict[str, str] | None) -> types.GetPromptResult:
    raise ValueError(f"Unknown prompt: {name}")

async def run():
    async with stdio_server() as (read_stream, write_stream):
        async with ServerSession(
            read_stream,
            write_stream,
            initialization_options=types.InitializationOptions(
                server_name="my-server",
                server_version="1.0.0",
                capabilities=server.get_capabilities(),
            ),
        ) as session:
            await session.run()

if __name__ == "__main__":
    import asyncio

    asyncio.run(run())

Steps to Create a Server that Provides File Resources

Step 1: Install Dependencies

pip install mcp-sdk

Step 2: Create Server Code

from mcp import ServerSession, StdioServerParameters, types
from mcp.server.stdio import stdio_server

async def handle_list_prompts() -> list[types.Prompt]:
    return []

async def handle_get_prompt(name: str, arguments: dict[str, str] | None) -> types.GetPromptResult:
    raise ValueError(f"Unknown prompt: {name}")

async def run():
    async with stdio_server() as (read_stream, write_stream):
        async with ServerSession(
            read_stream,
            write_stream,
            initialization_options=types.InitializationOptions(
                server_name="my-server",
                server_version="1.0.0",
                capabilities=server.get_capabilities(),
            ),
        ) as session:
            await session.run()

if __name__ == "__main__":
    import asyncio

    asyncio.run(run())

Step 3: Run the Server

python your_server_script.py

Example: Writing an MCP Client

Creating a Simple MCP Client

Here is a simple example of an MCP client:

from mcp import ClientSession, StdioClientParameters, types
from mcp.client.stdio import stdio_client

async def run():
    async with stdio_client() as (read_stream, write_stream):
        async with ClientSession(
            read_stream,
            write_stream,
            connection_options=types.ConnectionOptions(),
        ) as session:
            await session.run()

if __name__ == "__main__":
    import asyncio

    asyncio.run(run())

🤝 Contribution Guidelines

How to Contribute Code

1. Fork the project repository.
2. Create a feature branch.
3. Commit your code changes.
4. Create a Pull Request.

📄 License

This project is licensed under the Apache 2.0 license.