T22: Full Stack AI App

This is where everything comes together. A full-stack application connects the frontend (what users see) to the backend (server logic and data) to the AI layer (intelligent responses). Think of it as building a complete restaurant - the dining room, the kitchen, and the chef all working in harmony.

Architecture Overview

The frontend sends user input to your Node.js server. The server manages sessions, validates data, and forwards prompts to Ollama. Responses flow back through the same chain.

Connecting the Layers

// Server: Bridge between frontend and AI
app.post("/api/chat", authenticate, async (req, res) => {
    const { message } = req.body;
    const userId = req.session.userId;

    // Save to database
    db.prepare("INSERT INTO messages (user_id, role, content) VALUES (?, ?, ?)")
      .run(userId, "user", message);

    // Get conversation history
    const history = db.prepare("SELECT role, content FROM messages WHERE user_id = ? ORDER BY id")
      .all(userId);

    // Call Ollama
    const aiResponse = await chat(history);

    // Save AI response
    db.prepare("INSERT INTO messages (user_id, role, content) VALUES (?, ?, ?)")
      .run(userId, "assistant", aiResponse);

    res.json({ reply: aiResponse });
});