code-society-25-2

Supplemental 5: Agentic AI (Slides)

Pre-work

Please review the following resources before lecture:

Tools

Demo: MCP Server Implementation

The fullstack_demo project includes a working Model Context Protocol (MCP) server that demonstrates how AI agents can interact with a real application.

What’s Included

The demo shows how to:

Expose application functionality (todos) through MCP tools
Implement authentication for AI agents using Clerk
Support multiple transport types (stdio, SSE, HTTP)
Deploy an MCP server to production (Vercel)

Code Location

📂 MCP Server Route: lib/javascript/fullstack_demo/src/app/[transport]/route.ts

This file demonstrates:

Creating MCP tools with createMcpHandler
Defining tool schemas with Zod validation
Implementing authentication with withMcpAuth
Handling different transport mechanisms

Quick Start

cd lib/javascript/fullstack_demo
npm install
npm run dev

# In another terminal, connect an AI agent
npx -y mcp-remote http://localhost:3000/mcp

Full Documentation

For detailed setup instructions and architecture details, see:

This site is open source. Improve this page.