Hook to expose committed TTS audio before playing back to user

[BaiMoHan]

I'm looking for an elegant way to save the audio output of the Text-to-Speech (TTS) system, specifically the audio that is played back to users during interactions, as well as the audio recorded from users speaking. Having these audio files saved would greatly help in reconstructing the live experience.

Is there a method or existing implementation that allows for automatic saving of this audio data? Additionally, does LiveKit provide any built-in functionality for automatically saving these audio recordings?

Any guidance or examples would be greatly appreciated!

I use the LiveKit Cloud, but there's nothing in my Egress web console. But the docs told me that Egress is ready to use without additional configuration for cloud customers.