Test Remote

You can run a FastMCP server locally to expose programmable tools and then access them from an MCP client. This setup uses the uv runtime to initialize and manage a FastMCP server written in Python, including basic example tools you can test end-to-end.

How to use

Start the server using the local runtime and then test it with an MCP inspector client. You will be able to browse available tools, invoke them, and see results in real time.

How to install

Prerequisites you need before starting:

• Install the uv runtime on your system.

• Ensure Python is installed and available on your PATH.

Step-by-step commands to set up a new MCP server project and run it locally:

uv init
uv add fastmcp
# Create your server entry point, e.g., main.py
# Write your FastMCP server code in main.py
# Start the server for testing
uv run fastmcp run main.py
# Optional: test with the MCP Inspector in your browser
# When ready, you can install a local desktop integration if needed

Additional sections

This guide includes a minimal, working server example and steps to deploy or extend it. You’ll learn how to test the server locally, connect with an MCP client, and push your project to version control as you build your MCP capabilities.

import random
from fastmcp import FastMCP

# Create a FastMCP server instance
mcp = FastMCP(name="Demo Server")

@mcp.tool
def roll_dice(n_dice: int=1) -> list[int]:
    """Roll n_dice 6 sided dice and return the results."""
    return [random.randint(1,6) for _ in range(n_dice)]

@mcp.toll
def add_numbers(a: float, b: float) -> float:
    """Add two numbers together."""
    return a + b

if __name___ == "__main__":
    mcp.run()

Available tools

roll_dice

Rolls a specified number of six-sided dice and returns the results as a list.

add_numbers

Adds two numbers and returns the sum as a float.