Building AI Applications: First AI Integration

Part 3: First AI Integration

Recap: Parts 1-2 established our foundation and created a basic Node.js project structure.

Today we make the exciting leap to an AI-powered application by connecting to Azure OpenAI Service.

Azure OpenAI Setup

Request access at portal.azure.com (takes 1-2 days approval). Once approved, create an OpenAI resource and note your endpoint, API key, and deployment name.

Add AI Dependencies

npm install @azure/openai

Update .env:

AZURE_OPENAI_API_KEY=your_key
AZURE_OPENAI_ENDPOINT=https://your-resource.openai.azure.com
AZURE_OPENAI_DEPLOYMENT=your_deployment

Create AI Service

Create src/services/aiService.js:

const { OpenAIApi, Configuration } = require('@azure/openai');

const client = new OpenAIApi(new Configuration({
  apiKey: process.env.AZURE_OPENAI_API_KEY,
  basePath: `${process.env.AZURE_OPENAI_ENDPOINT}/openai/deployments/${process.env.AZURE_OPENAI_DEPLOYMENT}`,
  baseOptions: {
    headers: { 'api-key': process.env.AZURE_OPENAI_API_KEY },
    params: { 'api-version': '2023-12-01-preview' }
  }
}));

async function getAIResponse(message) {
  const response = await client.createChatCompletion({
    model: process.env.AZURE_OPENAI_DEPLOYMENT,
    messages: [
      { role: 'system', content: 'You are a helpful customer support assistant.' },
      { role: 'user', content: message }
    ]
  });
  return response.data.choices[0].message.content;
}

module.exports = { getAIResponse };

Add Chat Endpoint

Update src/app.js:

require('dotenv').config();
const express = require('express');
const { getAIResponse } = require('./services/aiService');

const app = express();
app.use(express.json());

app.post('/chat', async (req, res) => {
  const aiResponse = await getAIResponse(req.body.message);
  res.json({ response: aiResponse });
});

app.listen(3000, () => console.log('AI server running'));

Test It

curl -X POST http://localhost:3000/chat \
  -H "Content-Type: application/json" \
  -d '{"message": "Hi, I need help"}'

You now have an AI-powered application! In Part 4, we’ll add LangChain for conversation memory.