Input types for CUSTOM and FULL modes

Hi Team,

I would like to clarify the following. Based on the docs, in FULL mode, I can use avatar.speak_text and avatar.speak_response where inputs are texts, while in CUSTOM mode, I can only use agent.speak where the input is an audio.

Is it correct that in CUSTOM mode, text is not an acceptable input for my avatar to speak/lip-sync, only audio?

Thank you very much.