How to create assist automation to save your voice message?

Im very new, dont know where to start and how to ask. My ultimate want is i set up ESP box3 with mic, to work on wake word, then i say the trigger words “write down a message” and it sends th enext words im speaking to my matrix room.

I have made the matrix integration and i can send data to the room. Im struggling to understand how to make it so that the assist/ant sends my spoken data that is converted to speech-to-text in text form to my matrix room.

Is there a plugin/addon/integration/patch or something i could look into?

And the second part to this want is that when i trigger action from my matrix room, i want the text-to-speech send back the message to the ESP box device…

Can someone help?