Build Your Own MCP Server

0 %

Course content

Problem: You Don't Know Where to Start

Too Many Options, No Clear Path

Cutting through the noise to set up your first MCP server

You have decided to build an MCP server. But when you look at the documentation, you are immediately faced with choices: TypeScript or Python? stdio or HTTP? What SDK version? What project structure? It is easy to get lost before writing a single line of code.

Let us cut through the noise and get you set up in the simplest way possible.

Choosing Your Language

Feature	TypeScript	Python
SDK Package	@modelcontextprotocol/sdk	mcp[cli]
Best for	Web APIs, Node.js ecosystems	Data processing, scripting, Odoo
Type safety	Zod schemas (excellent)	Type hints + Pydantic
Quick start	npx @anthropic-ai/create-mcp	uvx create-mcp-server

Our Recommendation

If you are connecting to web APIs or building something from scratch, use TypeScript. If you are working with data, scripts, or Odoo (like in the VaryShop ecosystem), use Python.

TypeScript Setup

Terminal

# Create project directory
mkdir my-mcp-server
cd my-mcp-server

# Initialize project
npm init -y
npm install @modelcontextprotocol/sdk zod
npm install -D typescript @types/node

# Create tsconfig.json
npx tsc --init --target ES2022 --module Node16 \
  --moduleResolution Node16 --outDir dist

Now create your server file src/index.ts:

src/index.ts

import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";
import { z } from "zod";

const server = new McpServer({
  name: "my-first-server",
  version: "1.0.0",
});

server.tool(
  "greet",
  "Greet a person by name",
  { name: z.string().describe("The person's name") },
  async ({ name }) => ({
    content: [{ type: "text", text: `Hello, ${name}!` }],
  })
);

async function main() {
  const transport = new StdioServerTransport();
  await server.connect(transport);
}
main();

Just 20 Lines

That is it — 20 lines of code for a working MCP server. It registers one tool called greet that takes a name and returns a greeting.

Python Setup

Terminal

# Create project
mkdir my-mcp-server
cd my-mcp-server

# Set up with uv (recommended)
uv init
uv add "mcp[cli]" httpx

server.py

from mcp.server.fastmcp import FastMCP

mcp = FastMCP("my-first-server")

@mcp.tool()
def greet(name: str) -> str:
    """Greet a person by name."""
    return f"Hello, {name}!"

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

Even More Concise

Python FastMCP is just 12 lines. The decorator @mcp.tool() automatically converts the function signature and docstring into a tool definition.

Transport: stdio vs Streamable HTTP

stdio (Standard I/O)

The server runs as a local process. Input/output goes through stdin/stdout. Simplest and most common transport for local development.

Streamable HTTP

The server runs as a web service. Communication over HTTP with server-sent events. Use this for remote or shared servers.

Start Simple

Start with stdio. It is simpler, requires no network setup, and is what Claude Code uses by default. You can always add HTTP later.

Your First Test

MCP Inspector

# TypeScript
npx @modelcontextprotocol/inspector node dist/index.js

# Python
mcp dev server.py

The inspector opens a web UI where you can see your registered tools, call them manually, and verify they work correctly.

Key Takeaway

Setting up an MCP server is surprisingly simple. In 20 lines of TypeScript or 12 lines of Python, you have a working server with a registered tool. The hard part is not the setup — it is deciding what tools to build. We will tackle that next.