Deepdub SDK

This is the official SDK for the Deepdub API. It provides a simple way to interact with the API and generate dubbing scripts for your videos.

Install

npm install --save @deepdub/node

# or

yarn add @deepdub/node

Usage

const { DeepdubClient } = require('@deepdub/node');

require('dotenv').config();

async function main() {
  const deepdub = new DeepdubClient('YOUR_API_KEY');
  const voicePromptId = 'YOUR_PROMPT_ID';
  await deepdub.connect();

  // You can generate into a buffer:
  //
  const buffer = await deepdub.generateToBuffer('Hello, this is a test.', {
    locale: 'en-US',
    voicePromptId,
  });

  // Or you can generate into a file:
  //
  await deepdub.generateToFile('./generated.wav', 'Hello, this is a test.', {
    locale: 'en-US',
    voicePromptId,
  });
}

main();

Client Initialization

By default, the DeepdubClient uses WebSocket protocol for streaming audio generation. You can also use HTTP protocol by passing the protocol option:

// Default: WebSocket protocol (supports streaming)
const deepdub = new DeepdubClient('YOUR_API_KEY');

// HTTP protocol (supports additional options like sampleRate and voiceReference)
const deepdub = new DeepdubClient('YOUR_API_KEY', { protocol: 'http' });

Protocol Comparison

| Feature | WebSocket | HTTP | | ---------------------------- | --------------- | ---- | | Streaming chunks (onChunk) | ✅ | ❌ | | sampleRate option | ⚠️ mp3 only | ✅ | | voiceReference option | ❌ | ✅ |

Use WebSocket (default) when you need real-time streaming. Use HTTP when you need voiceReference or sampleRate with non-mp3 formats.

Streaming Chunks

You can receive audio chunks as they arrive using the onChunk callback. This is useful for streaming audio playback or processing audio in real-time.

const buffer = await deepdub.generateToBuffer('Hello, this is a test.', {
  locale: 'en-US',
  voicePromptId: promptId,
  onChunk: (chunk) => {
    // Each chunk is a Buffer containing WAV audio data
    console.log(`Received ${chunk.length} bytes`);
  },
});

Headerless Chunks

By default, each chunk includes a WAV header. If you need raw audio data without headers (e.g., for streaming to an audio player), use the headerless option:

const buffer = await deepdub.generateToBuffer('Hello, this is a test.', {
  locale: 'en-US',
  voicePromptId: promptId,
  headerless: true,
  onChunk: (chunk) => {
    // Each chunk is raw PCM audio data (no WAV header)
    audioPlayer.write(chunk);
  },
});

The onChunk and headerless options are also available with generateToFile.