ALPANDIA
Back to Blog
Engineering15 minJune 18, 2025

Building Your First Voice AI Application with the ALPANDIA SDK

From npm install to live voice analysis in under 15 minutes. A step-by-step tutorial for getting started with the ALPANDIA TypeScript SDK.

D

Dev Kim

Developer Advocate

This tutorial walks you through building a complete voice AI application using the ALPANDIA TypeScript SDK — from installation to live transcription and translation in a single Node.js application.

Prerequisites

Installation

bash
npm install @alpandia/sdk

Initialize the Client

typescript
import Alpandia from '@alpandia/sdk';

const client = new Alpandia({
  apiKey: process.env.ALPANDIA_KEY,
  region: 'ap-southeast-1',
});

Start a Voice Stream

typescript
const session = await client.voice.stream({
  language: 'auto-detect',
  features: ['translate', 'sentiment', 'scam-detect'],
  onTranscript: (data) => {
    console.log(`[${data.speaker}]: ${data.text}`);
    console.log(`Translation: ${data.translation}`);
    console.log(`Sentiment: ${data.sentiment.label}`);
  },
  onScamAlert: (alert) => {
    if (alert.severity === 'high') {
      triggerSupervisorAlert(alert);
    }
  }
});

// Connect your audio source
await session.connect(audioInputStream);

That's all it takes to get live transcription, translation, and fraud detection running. The stream handles reconnection, buffering, and error recovery automatically.

Full API reference and additional examples are available in the ALPANDIA Developer Portal.

More in Engineering

Streaming vs Batch Transcription: Choosing the Right Architecture

11 min

Using Confidence Scores to Build Reliable Voice AI Pipelines

9 min

Integrating Real-Time Translation into Existing SIP Infrastructure

18 min