Voice CreationFrom Dataset

Node.js Workflow

Prerequisites

  • Resemble account with PRO plan or higher
  • API token (see Authentication)

1. Initialize the project

npm init -y
npm install @resemble/node

2. Configure the SDK

// resemble-clone-voice/index.js
import * as Resemble from "@resemble/node";
const apiKey = process.env.RESEMBLE_API_KEY;
if (!apiKey) {
  console.error("Please set the RESEMBLE_API_KEY environment variable.");
  process.exit(1);
}
Resemble.Resemble.setApiKey(apiKey);

3. Create the voice

async function createVoice() {
  const voiceName = "My Voice Clone";
  const dataset = "https://example.com/dataset.zip"; // Replace with your dataset URL
  console.log(`Submitting request to create voice: ${voiceName}`);
  try {
    const response = await Resemble.Resemble.v2.voices.create({
      name: voiceName,
      dataset_url: dataset,
    });
    if (response.success) {
      const voice = response.item;
      console.log(
        `Created voice ${voice.uuid}. Current status: ${voice.status}`
      );
    } else {
      console.error("Voice creation failed", response);
    }
  } catch (error) {
    console.error("Error creating voice:", error);
  }
}
createVoice();

Datasets must be publicly accessible and follow Resemble’s supported formats. Verify reachability with:

curl -I https://url/to/your/dataset.zip

4. Run the script

RESEMBLE_API_KEY=... node index.js

You’ll receive a response indicating the new voice UUID and the current training status (e.g., initializing). Monitor progress via the API or the Voices dashboard.

For a recordings-based workflow, see the custom voice with recordings guide.