Discussions

How can I have dynamic context?

When I start a LiveAvatar session, the avatar automatically says a default greeting like "Hi, I am from LiveAvatar..." I'm using the avatar in text-to-speech mode where my backend controls all responses via the repeat() method, so I need to disable this automatic greeting. I've tried calling interrupt() immediately after start() and setting voiceChat to false in SessionConfig, but the greeting still plays. Is there a configuration option or method to prevent the avatar from speaking this default greeting when the session starts?

Right now, if I want an avatar to repeat written text, the only available option is FULL mode, but that mode requires context as input. As a result, the avatar always generates an opening line, which I don’t want.

We want to explore using LiveAvatar, but we're not ready to commit developer time.

I am burning through credits during development/testing - there should be a test API/Avatar that has a lower cost for quick connection testing.

I would like to suggest an enhancement to the development environment. It would be nice to have a button in the context creation tool to have a gpt-style text-conversation with the context for debugging and finetuning. That would save developers from burning expensive credits (and ruining the ozone-layer) when testing. Also shouting and talking to a screen trying to upset the AI does really look stupid...

As we can fetch all available voices through below given api, but Kindly guide me how can we preview each one of them. How to check that how each of these voices actually sound so that we can take a decision which one to use ?