@sfgabe thanks for posting your entire HA config online. I still need to clean up mine & get mine posted likewise, and is on the list of things to do.
One thing I learned from your setup was the whole Janet speech engine. Ref URL for those who, like me, also missed that memo about Janet:
I am currently using Google TTS via all my remote Pi boxes via the speaker output. The core HA instance sends texts to say out to the Pi boxes via MQTT, and TTS conversion happens on each Pi. I’ve also been very surprised and pleased how “in sync” all the speakers sound across the house, despite different system loads on each Pi.
How is Janet working out for you? By any chance do you have some videos or audio recordings of how it sounds? Any implementation tips to watch out for? I think I may give this a try myself.
And again, thanks!