Docs
GitHub

Get Started

Documentation

Underscore is a generative audio SDK for interactive software. Describe a sound; get a programmable synth system your app can play, control, and adapt at runtime through our SDKs.

How It Works

  1. Generate — Describe your sound using natural language in the Underscore web app
  2. Publish — Make the composition available via API (unlisted or public)
  3. Integrate — Load the synth in your application using an SDK
  4. Control — Adjust parameters in real time as users interact

SDKs

PlatformStatusLink
Web (TypeScript/JS)AvailableWeb SDK Docs
Unity (C#)Coming soon
iOS (Swift)Coming soon

The HTTP API is currently intended to be consumed through these SDKs. A standalone HTTP reference is not published yet; see the underscore-web-sdk source on GitHub for how requests are shaped if you need to call it directly.

Curious what actually makes the sound? See Synthesis Engine.