Skip to content

Client

Model Context Protocol (MCP) is an open protocol that standardizes how applications provide context to LLMs.

This guide explains how to build MCP clients that connect to MCP servers.

Installation#

To integrate MCP tools into the AG2 framework, install the required dependencies:

pip install -U ag2[openai,mcp]

Note: If you have been using autogen or pyautogen, all you need to do is upgrade it using: 

pip install -U autogen[openai,mcp]
or 
pip install -U pyautogen[openai,mcp]
as pyautogen, autogen, and ag2 are aliases for the same PyPI package.

Imports#

from pathlib import Path

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

from autogen.agentchat import AssistantAgent
from autogen.mcp import create_toolkit

# Only needed for Jupyter notebooks
import nest_asyncio

nest_asyncio.apply()

Setting Up an MCP Server#

We will create a simple MCP server that exposes two tools: add and multiply.

For more details on creating MCP servers, visit the MCP Python SDK documentation.

Let's create a Python script (math_server.py) with the following content:

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

# Write content to a file
math_server_path = Path("math_server.py")
math_server_path.write_text(math_server_file_content)

Connecting AG2 to the MCP Server#

To connect AG2 to the MCP server, follow these steps: - Set up connection parameters – Define how AG2 will communicate with the MCP server. - Start a client session – This session handles communication between AG2 and the server. - Wrap MCP tools into a toolkit – This makes them accessible to AG2. - Register the toolkit with an AG2 agent – This enables AG2 to use MCP tools. - Start conversation - Send a message with desired goal (e.g. adding two numbers).

# Create server parameters for stdio connection
server_params = StdioServerParameters(
    command="python",  # Executable
    args=[str(math_server_path)],  # Optional command line arguments
)

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

    # Create a toolkit with available MCP tools
    toolkit = await create_toolkit(session=session)
    agent = AssistantAgent(name="assistant", llm_config={"model": "gpt-4o-mini", "api_type": "openai"})
    # Register MCP tools with the agent
    toolkit.register_for_llm(agent)

    # Make a request using the MCP tool
    result = await agent.a_run(
        message="Add 123223 and 456789",
        tools=toolkit.tools,
        max_turns=2,
        user_input=False,
    )
    await result.process()

Expected Output#

user (to assistant):

Add 123223 and 456789

--------------------------------------------------------------------------------
assistant (to user):

***** Suggested tool call (call_kO3xnTqJyOxTYaJCdyS1VOKw): add *****
Arguments:
{"a": 123223, "b": 456789}
********************************************************************

--------------------------------------------------------------------------------

>>>>>>>> EXECUTING FUNCTION add...
Call ID: call_kO3xnTqJyOxTYaJCdyS1VOKw
Input arguments: {'a': 123223, 'b': 456789}
user (to assistant):

***** Response from calling tool (call_kO3xnTqJyOxTYaJCdyS1VOKw) *****
('580012', None)
**********************************************************************

--------------------------------------------------------------------------------
assistant (to user):

The sum of 123223 and 456789 is 580012.

TERMINATE