sarvavox.Try Sarvavox

DEVELOPER QUICKSTART

Give your AI
a global voice.

Start with the public sample catalog. Inference API keys and official SDKs are not yet self-service; contact the team to discuss an integration.

Your first sample.

  1. Open the playground and choose a voice.
  2. Fetch the JSON catalog from the same origin. Each item includes a language code, transcript, translation, and audio path.
  3. Play the chosen sample from a user click. No API key or package installation is needed.

These examples retrieve saved audio. They do not synthesize new text.

JavaScript · sample playback
const samples = await fetch('/demos.json')
  .then(response => response.json());

const hindi = samples.find(s => s.language === 'hi');
const audio = new Audio(hindi.audio);
// Call from a user interaction, such as a click.
await audio.play();

Runs on sarvavox.com or the local website. Saved audio, no inference.

Download a sample catalog.

curl --fail https://sarvavox.com/demos.json -o demos.json

The catalog works with standard HTTP clients. On another origin, download the catalog and audio server-side; cross-origin browser access is not guaranteed.

Run the website locally.

Requires Node.js 20+ and access to the repository. There are no npm dependencies to install.

git clone https://github.com/TwoMuchLabs/sarvavox.git
cd sarvavox
npm run dev
# Open http://127.0.0.1:4173
View sourceDiscuss API access