A big problem for me is, that Android is not allowing custom DoT servers. Even though the system supports DoH and is even using it for their built in resolver (Google/cloud flare)
Networks that only whitelist TCP 433 (some guest wifis) will fail to use DoT.
I have Kodi running in a docker container with gui (even HDR support) on my otherwise "headless" server.
Kodi just has the most flexibility because of it's plugins.
You should be able to follow the devs here https://github.com/esphome/home-assistant-voice-pe