Demo MCP Server

Creating a Python MCP Server: A Step-by-Step Guide

The Model Context Protocol (MCP) provides a standardized approach for connecting context sources to large language models (LLMs). This tutorial demonstrates how to build a functional MCP server using Python's MCP SDK, enabling you to expose data, tools, and templates to LLM applications.

What is MCP?

MCP creates a bridge between LLM applications and external context sources. It allows you to modularize different aspects of LLM interactions:

Data provision through resources (similar to read-only endpoints)
Action execution via tools (comparable to API functions)
Template management using prompts for reusable interactions

Core MCP Components

MCP servers implement three fundamental building blocks, each serving different purposes:

Component	Controlled By	Purpose	Common Uses
Prompts	User	Interactive templates triggered by user selection	Command shortcuts, menu items
Resources	Application	Data managed by the client for LLM context	File content, API data
Tools	LLM	Functions the model can execute independently	Calculations, API calls, data modifications

Understanding these distinctions helps you design effective MCP servers that properly separate concerns.

Server Feature Advertising

MCP servers announce their capabilities during startup, allowing clients to adapt their behavior:

Feature	Configuration Flag	What It Enables
prompts	`listChanged`	Dynamic prompt template updates
resources	`subscribe`, `listChanged`	Data exposure with live updates
tools	`listChanged`	Function discovery and execution
logging	Default	Debug output configuration
completion	Default	Argument suggestion support

Getting Started

Requirements

Ensure your environment includes:

Python 3.7+ (Python 3.11+ recommended)
pip package manager
Node.js 18.x

Installation Options

Choose one of these installation methods:

Standard pip installation:

pip install "mcp[cli]"

Using uv (recommended for project management):

uv init mcp-server
cd mcp-server
uv add "mcp[cli]"

Project Structure

Organize your project as follows:

mcp-server/
├── server.py
├── pyproject.toml (if using uv)
└── README.md

Implementation

Building Your Server

Create server.py with the following foundation:

# server.py
from mcp.server.fastmcp import FastMCP

# Initialize the MCP server
mcp = FastMCP("Demo Server")

# Define a calculation tool
@mcp.tool()
def add(a: int, b: int) -> int:
    """
    Performs addition of two integers.
    
    Args:
        a: First number
        b: Second number
    
    Returns:
        Sum of both numbers
    """
    return a + b

# Create a dynamic resource
@mcp.resource("greeting://{name}")
def get_greeting(name: str) -> str:
    """
    Generates a personalized greeting.
    
    Args:
        name: Person's name for the greeting
    
    Returns:
        Formatted greeting message
    """
    return f"Hello, {name}!"

# Add a prompt template
@mcp.prompt()
def review_code(code: str) -> str:
    """
    Creates a code review template.
    
    Args:
        code: Source code to review
    
    Returns:
        Formatted review prompt
    """
    return f"Please review this code:\n\n{code}"

# Server execution
if __name__ == "__main__":
    mcp.run()

Testing Your Server

Using the Development Inspector

The most efficient way to test your server is through the built-in development interface:

mcp dev server.py

This command:

Launches your server with live reloading
Opens a web interface at http://localhost:6274/
Provides interactive testing capabilities

Inspector Configuration

In the Inspector interface, configure the transport settings:

Transport Type: STDIO  
Command: python  
Arguments: server.py

Click Connect to establish the connection.

Testing Each Component

Tool Testing

Navigate to the Tools section
Select the add tool
Input test values (e.g., a = 10, b = 15)
Execute and verify the result (25)

Resource Testing

Go to Resources → Resource Templates
Select get_greeting
Enter a name (e.g., Alice)
Click Read Resource
Verify the response: "Hello, Alice!"

Prompt Testing

Access Prompts → List prompts
Select review_code
Input sample code: print(1+1)
Execute to see the formatted prompt output

Run this in a separate terminal while your server is active.

Understanding Server Behavior

When you run python server.py directly, the server appears inactive because it uses stdio transport and waits for client connections. This is normal behavior - the server needs a client (like the Inspector or your custom client) to interact with it.

This server cannot be installed

security - not tested

license - not found

quality - not tested

How are these scores calculated?

A simple MCP server that provides a calculator tool, dynamic greeting resources, and code review prompts, demonstrating how to build standardized interfaces for LLM applications.

Related MCP Servers

Claude-LMStudio Bridge
infinitimeless
-
security
F
license
-
quality
An MCP server that allows Claude to interact with local LLMs running in LM Studio, providing access to list models, generate text, and use chat completions through local models.
Last updated -
8
Python
Vibe-Coder MCP Server
crazyrabbitLTC
A
security
F
license
A
quality
An MCP server that implements a structured workflow for LLM-based coding, guiding development through feature clarification, documentation generation, phased implementation, and progress tracking.
Last updated -
10
14
TypeScript
HostBridge MCP Server
elblanco2
-
security
A
license
-
quality
An MCP server that helps novice developers deploy web applications through conversational interfaces, bridging the gap between LLMs and various hosting environments.
Last updated -
1
Python
MIT License
Moodle MCP Server
peancor
A
security
A
license
A
quality
An MCP server that enables LLMs to interact with Moodle platforms to manage courses, students, assignments, and quizzes through natural language commands.
Last updated -
7
13
JavaScript
MIT License

View all related MCP servers