Skip to main content

​
Overview

We expect you to work with API in the following way:
  1. You create a session on your backend and send a session token to the frontend.
  2. You use our Web SDK there to control the avatar speech.
  3. You terminate a session in the backend if needed. The session will be automatically terminated when user leaves.
The current recommended approach to the integration is to use OpenAI Realtime conversation engine.