Step 1. Create the MCP server configuration

1Create a server.py file that instantiates a FastMCP server.
For example, add the following code:
from mcp.server.fastmcp import FastMCP
mcp = FastMCP("My API Server")
2Define at least one MCP tool.
The following sample tool takes the prompt and conversation ID as parameters, posts them through the REST API, and returns the AI agent's response if successful:
@mcp.tool()
def interact_with_agent(prompt: str, conversationID: str) -> str:

"""Uses the information in the prompt to interact with the AI agent."""

# Define authentication
import requests
headers = {
"Authorization": "<authorization string>"
}

# Define the API request
url = "<Endpoint URL from AI Agent Engineering>"

#Create the data object required by the API
data = {
"prompt": prompt,
"conversationID": conversationID
}

# Post the API request
import sys
response = requests.post(url, json=data, headers=headers)

# Handle the response to extract the required payload
if response.status_code == 200:
return str(response.json())

#Add error handling
return None
For more information about creating tools with FastMCP, see "Tools" in the FastMCP documentation.