Building a Minimal MCP Server

One Protocol, Many Tools

An MCP server exposes tools over stdio or HTTP. Any MCP-capable client (Claude Desktop, Cursor, Claude Code) can call them. Build once, use everywhere.

npm install @modelcontextprotocol/sdk zodOfficial SDK plus Zod for schemas.

// server.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: "word-tools", version: "0.1.0", }); server.tool( "word_count", "Count words in a string.", { text: z.string().min(1) }, async ({ text }) => { const count = text.trim().split(/\s+/).length; return { content: [{ type: "text", text: String(count) }], }; } ); server.tool( "reverse", "Reverse a string.", { text: z.string() }, async ({ text }) => ({ content: [{ type: "text", text: [text].reverse().join("") }], }) ); const transport = new StdioServerTransport(); await server.connect(transport);Two tools, typed inputs, string outputs. Run with `node server.js` from an MCP-capable client config.

// claude_desktop_config.json snippet { "mcpServers": { "word-tools": { "command": "node", "args": ["/absolute/path/to/server.js"] } } }Point Claude Desktop at your script. Restart, and the tools appear.

Understanding "Building a Minimal MCP Server" in practice: AI-assisted coding shifts work from syntax recall to design thinking — models handle boilerplate so you focus on architecture. Model Context Protocol lets agents plug into your tools. A 40-line server exposes a real capability to Claude — and knowing how to apply this gives you a concrete advantage.

• Apply MCP in your ai-coding workflow to get better results
• Apply tools in your ai-coding workflow to get better results
• Apply stdio transport in your ai-coding workflow to get better results
• Apply schema in your ai-coding workflow to get better results

1. Use AI to generate unit tests for an existing function
2. Ask AI to refactor a messy function and explain the changes
3. Have AI suggest a code review for a recent pull request

The big idea: MCP is the USB port for agents. Write a tiny server once, and every MCP client can use your tools.