Anthropic官方推荐！LangChain MCP双协议支持全球800+工具

要点

• 适配了Anthropic Model Context Protocol (MCP) 所有的工具
• 既可以进行本地通信stdio，也可以及远程通信SSE
• 与LangChain LangGraph生态无缝适配

LangChain MCP 实战

pip install langchain-mcp-adapters

1、 首先创建MCP服务

• 这里创建了一个数学计算服务math_server.py,mcp服务的名称叫Math,采用的通信方式为stdio

# math_server.py
from mcp.server.fastmcp import FastMCP

mcp = FastMCP("Math")

@mcp.tool()
def add(a: int, b: int) -> int:
    """Add two numbers"""
    return a + b

@mcp.tool()
def multiply(a: int, b: int) -> int:
    """Multiply two numbers"""
    return a * b

if __name__ == "__main__":
    # transport表示mcp通信的类型：stdio（标准输入输出） 适用于本地通信；SSE（Server-Sent Events） 适用于远程通信 
    mcp.run(transport="stdio")

2、创建client连接MCP服务

from mcp import ClientSession, StdioServerParameters
from mcp.client.stdio import stdio_client

from langchain_mcp_adapters.tools import load_mcp_tools
from langgraph.prebuilt import create_react_agent

from langchain_openai import ChatOpenAI
model = ChatOpenAI(model="gpt-4o")

server_params = StdioServerParameters(
    command="python",                  # 用python命令执行本地通信
    args=["/path/to/math_server.py"],  # 本地地址必须是绝对路径
)

asyncwith stdio_client(server_params) as (read, write):
    asyncwith ClientSession(read, write) as session:
        # 初始化链接client session
        await session.initialize()

        # 加载mcp服务中的工具,注册为langchian中的工具
        tools = await load_mcp_tools(session)

        # langchian创建agent，调用mcp中的工具进行运算
        agent = create_react_agent(model, tools)
        agent_response = await agent.ainvoke({"messages": "what's (3 + 5) x 12?"})

3、 创建多个MCP服务

• 创建一个查询天气的服务math_server.py

# math_server.py
...

# weather_server.py
from typing import List
from mcp.server.fastmcp import FastMCP

mcp = FastMCP("Weather")

@mcp.tool()
asyncdef get_weather(location: str) -> str:
    """Get weather for location."""
    return"It's always sunny in New York"

if __name__ == "__main__":
    mcp.run(transport="sse")  # HTTP with SSE（Server-Sent Events） 适用于远程通信 

启动远程服务

python weather_server.py

4、 连接多个MCP服务

from langchain_mcp_adapters.client import MultiServerMCPClient
from langgraph.prebuilt import create_react_agent

from langchain_openai import ChatOpenAI
model = ChatOpenAI(model="gpt-4o")

asyncwith MultiServerMCPClient(
    {
        "math": {  # 启动math服务
            "command": "python",
            "args": ["/path/to/math_server.py"],
            "transport": "stdio",
        },
        "weather": {  # 启动weather服务
            # make sure you start your weather server on port 8000
            "url": "http://localhost:8000/sse",
            "transport": "sse",
        }
    }
) as client:
    agent = create_react_agent(model, client.get_tools())
    math_response = await agent.ainvoke({"messages": "what's (3 + 5) x 12?"})
    weather_response = await agent.ainvoke({"messages": "what is the weather in nyc?"})

https://github.com/langchain-ai/langchain-mcp-adapters

文章来源：微信公众号-CourseAI，原始发表时间：2025年03月30日。