https://modelcontextprotocol.io/introduction

快速搭建

服务端

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__":  
    mcp.run(transport="stdio")

可视化server(需要安装npx,我这里装过了，开了代理下的包@modelcontextprotocol/inspector） mcp dev server.py

mcp

客户端

我这里用的langchain框架langchain_mcp_adapters

client.py

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 model import llm

server_params = StdioServerParameters(
    command="python",
    # Make sure to update to the full absolute path to your math_server.py file
    args=["simple_mcp_server.py"],
)


async def run():
    async with stdio_client(server_params) as (read, write):
        async with ClientSession(read, write) as session:
            # Initialize the connection
            await session.initialize()

            # Get tools
            tools = await load_mcp_tools(session)
            print(tools)

            # Create and run the agent
            agent = create_react_agent(llm, tools)
            agent_response = await agent.ainvoke({"messages": "what's (3 + 5) x 12?"})
            print(agent_response)


if __name__ == "__main__":
    import asyncio

    asyncio.run(run())

model.py

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    api_key='sk_xxxx',
    model='deepseek/deepseek-v3-0324',
    base_url='https://xxx.com/v3/openai'
)