Custom Python MCP Server

Guide for creating and using custom Python MCP servers with PraisonAI agents

Custom Python MCP Server

Quick Start

1. Create MCP Server

   Create a new file app.py with your custom MCP server implementation:

   import yfinance as yf
   from mcp.server.fastmcp import FastMCP
   
   mcp = FastMCP("stock_prices")
   
   @mcp.tool()
   async def get_stock_price(ticker: str) -> str:
       """Get the current stock price for a given ticker symbol.
   
       Args:
           ticker: Stock ticker symbol (e.g., AAPL, MSFT, GOOG)
   
       Returns:
           Current stock price as a string
       """
       if not ticker:
           return "No ticker provided"
       try:
           stock = yf.Ticker(ticker)
           info = stock.info
           current_price = info.get('currentPrice') or info.get('regularMarketPrice')
           if not current_price:
               return f"Could not retrieve price for {ticker}"
           return f"${current_price:.2f}"
   
       except Exception as e:
           return f"Error: {str(e)}"
   
   if __name__ == "__main__":
       mcp.run(transport='stdio')

2. Install Dependencies

   Install the required dependencies in a conda environment:

   pip install yfinance mcp

3. Create Agent Integration

   Create a new file stock_agent.py with the following code:

   from praisonaiagents import Agent, MCP
   
   agent = Agent(
       instructions="""You are a helpful assistant that can check stock prices and perform other tasks.
       Use the available tools when relevant to answer user questions.""",
       llm="gpt-4o-mini",
       tools = MCP("/path/to/python /path/to/app.py")
   )
   
   # NOTE: Replace with your actual Python path and app.py file path
   
   agent.start("What is the stock price of Tesla?")

4. Run the Agent

   Execute your script:

   python stock_agent.py

Features

Implementation Details

FastMCP Class

The FastMCP class from the mcp-python-sdk package provides a simple way to create MCP servers in Python:

from mcp.server.fastmcp import FastMCP

# Create an MCP server with a name
mcp = FastMCP("my_tools")

# Define a tool using the @mcp.tool decorator
@mcp.tool()
async def my_tool(param1: str, param2: int) -> str:
    """Tool description with clear documentation.

    Args:
        param1: Description of param1
        param2: Description of param2

    Returns:
        Description of the return value
    """
    # Tool implementation
    return f"Processed {param1} with {param2}"

# Run the server with stdio transport
if __name__ == "__main__":
    mcp.run(transport='stdio')

Agent Integration

To use your custom MCP server with PraisonAI agents, use the MCP class to specify the command to run your Python script:

from praisonaiagents import Agent, MCP

agent = Agent(
    instructions="Agent instructions",
    llm="gpt-4o-mini",
    tools=MCP(
        command="python",  # Or full path to Python
        args=["path/to/your/mcp_server.py"]  # Path to your MCP server script
    )
)