Your Sonora trial is active,
get started in 5 minutes

Welcome to Sonora! Your free trial is now active.

The Sonora Diarization API helps you build more accurate speaker diarization pipelines. It segments recorded audio into speaker turns, answering a simple question: who spoke when?

Clean diarization improves:

• Speech-to-text accuracy: fewer word errors from speaker confusion.
• Analytics: reliable speaker turns for insights and dashboards.
• Training data: better-labeled datasets for downstream models.

1. Create an API key

Create your key under account settings. Keys are shown only once, so copy yours and store it securely.

2. Diarize your first file

Send a POST to /v1/diarize with your API key as a bearer token. In the JSON body, set audio_url to your recorded audio’s URL and speakers to "auto" to detect the speaker count automatically.

POST /v1/diarize
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

{
  "audio_url": "YOUR_AUDIO_URL",
  "speakers": "auto"
}

Replace YOUR_API_KEY and YOUR_AUDIO_URL with your values. The response lists speaker turns with start, end, and speaker fields.

Follow the quickstart guide for the full walkthrough.

3. Try the playground

Run a sample file in the playground. It uses the same model, with no code required.

We’re looking forward to seeing what you build.

The Sonora team