How to handle the input independently

Hi, I'm using the streaming function, the reason is I want to use the interactive avatar to make the interaction more smooth.

But after get the audio track from browser, if I still use SDK, is there any way I can get the transcribed text but not using the automatic generated answer? That means after I get user input in text, can I generate the answer in my own code and manually call speak method of the avatar?

Thanks!