Separate processing server

I have HASS setup on my raspberry pi 3B+, and recently I have added voice support.

The only problem is that the 3B+ chugs trying to handle processing.

I want to improve this by having a separate server handle STT and TTS (Which I have and it is quite beefy).

But I do not see any documentation to have a separate server handle the processing? Would this need a separate PR?