Discussions

When using API from a webpage in full mode, we let the webpage respond on the conversation. So for example when the avatar says "red", an object on the page changes it's color to red. Currently, I only get the avatar transcript AFTER the avatar stops speaking. But I assume your backend already knows what the avatar is going to say, driven by the LLM.

Long story short: can we get the transcript BEFORE or at the same time the avatar starts speaking. We can then make the response mechanism much more smooth