๐๐จ๐๐๐ ๐ ๐ง๐๐ก๐ฌ ๐ ๐๐ฃ ๐ฆ๐๐ฅ๐ฉ๐๐ฅ ๐๐ก ๐๐๐ฉ๐๐ฆ๐๐ฅ๐๐ฃ๐ง You can build a tiny server that exposes tools to LLMs using the Model Context Protocol (MCP). This guide shows you how to create a minimal, runnable example.
Here's what you get:
- A working server that exposes a single tool: get-repo-stats
- Quick instructions to run locally or remotely
- Short adapter patterns for OpenAI, Google Gemini, and Anthropic Claude
To get started:
- Use Node 18+
- Run npm install
- Add "type": "module" to package.json for ESM
- Run node index.js
You can use MCP to extend models with live data, such as notifications, GitHub lookups, and database queries.
Source: https://dev.to/gaurav101/build-a-tiny-mcp-server-in-javascript-claude-codex-friendly-pbh Optional learning community: https://github.com/gaurav101/ai-experiment