Right now if you want voice synthesis (TTS) from Home Assistant your choices are Mary TTS which sounds like something from the 80’s, or cloud based spyware such as Google TTS.
Mimic 3 is a private, FOSS, local only (no cloud!) TTS option. It’s already compatible with Mary TTS, so HA can already talk to a Mimic server. There’s just no easy way to install it in Home Assistant.
An add-on would do just that! 
Or use Rhasspy’s TTS feature and just leave the other features unused.
Does installing Rhasspy really install Mimic 3 and make it available to HA? I thought Rhasspy only included Larynx, the precursor to Mimic 3…
The Rhasspy developer (Now mainly working at Mycroft) has made progress with the Mimic3 and have released docker images and debian packages.
Rhasspy seems to be able to be set up to use these.
Read the kaykoch comment here: Mimic 3 Text to Speech Launch! - #8 by Jarvy - Announcements - Rhasspy Voice Assistant