MCP (Model Context Protocol)

The Tool Integration Problem

Before MCP, every AI application had to build custom integrations for each tool. If you had 10 tools and 5 AI platforms, you built 50 connectors. Model Context Protocol (MCP), created by Anthropic, standardizes this into a universal protocol — like USB-C for AI.

MCP Architecture

Key Concepts

• MCP Host: The AI application (Claude Desktop, VS Code, your app) that needs to access tools.
• MCP Client: Lives inside the host. Maintains a 1:1 connection with each MCP server.
• MCP Server: Lightweight service that exposes tools, resources (data), and prompts via the MCP protocol.
• Transport: Servers connect via stdio (local) or SSE/HTTP (remote).

What MCP Servers Expose

• Tools: Functions the LLM can call (e.g., query_database, create_issue)
• Resources: Read-only data the LLM can access (e.g., file contents, DB schemas)
• Prompts: Reusable prompt templates that the host can surface to users

1# pip install mcp
2from mcp.server.fastmcp import FastMCP
3import sqlite3
4
5# Create an MCP server
6mcp = FastMCP("My Database Server")
7
8# Expose a tool
9@mcp.tool()
10def query_users(name: str) -> str:
11    \"\"\"Search users by name in the database.\"\"\"
12    conn = sqlite3.connect("app.db")
13    cursor = conn.execute(
14        "SELECT * FROM users WHERE name LIKE ?", (f"%{name}%",)
15    )
16    results = cursor.fetchall()
17    conn.close()
18    return str(results)
19
20# Expose a resource (read-only data)
21@mcp.resource("schema://database")
22def get_schema() -> str:
23    \"\"\"Return the database schema for context.\"\"\"
24    return "CREATE TABLE users (id INT, name TEXT, email TEXT)"
25
26# Run the server
27if __name__ == "__main__":
28    mcp.run(transport="stdio")  # Local: stdio, Remote: sse