Always-on Voice Sentence Detection

Is there a way to use either an ESP32 device, or a wyoming satellite, for always-on sentence detection in HA? My use case is: I’m alone in a silent room and I don’t even want to have to say a wake word because the only time I’d speak would be when I want HA to recognize a sentence.

I read through a few things but I couldn’t figure out any specifics here. Sorry if I missed this somewhere obvious!

I had a semi similar idea. I’m often the only one left in the lounge post 10pm, so thought about training ‘good night’ or ‘bed time’ as wake words and swapping one in at that time (swapping back out before the morning) You’d need to be using local wakeword on the satellite as I don’t know how to automate swapping it on the HA server itself.

However. I’ve not got around to it yet :smiley:

1 Like

Let me know how it goes if you do! My use case is a bit more dynamic, but if it’s possible I’ll update here (or link to a post in the contest section if I get the functionality up to snuff)