Is there some trick to getting this installed? My HAOS instance (x64 generic) has no updates pending despite refreshing by clicking “Check for updates” in the android app…
I’m not seeing it either (yet), and I’m on Haas-os.
Edit: nevermind, manually checking for the update in settings/system/updates made it show up
I actually had to force it via the CLI.
ha core updateThat didn’t work either, I had to manually specify the version:
ha core update --version 2025.7.0
@HiTekRedNek @thehatfox My instance found the update 3 hours ago.
The voice focus is fucking el-el-lame
Removing dependency on using cloud voice like Google or Amazon is awesome. I wouldn’t use voice if that’s how I had to do it, and I really like voice control for my HA ecosystem.
To me it’s very gimmicky and requires a lot of work. It probably fits English speakers better as I’ve not had any luck using local voice.
You do need a bit of horsepower to run Whisper locally. A Nabu Casa subscription gets you their voice processing, and it’s quick. Granted, it’s not local but I trust them a pile more than I trust either of the other two, and if I had to I’d buy a GPU and run it locally, that’s their focus anyway.
Agree. I do not speak to the machine. It does not speak to me. It is a happy arrangement.
I can see the use as an accessibility aid, but nothing more.
All I have ever wanted from a machine is to be able to say, “I’m busy right now, but I’ve had a thought; here - hold this for me”…
…without it telling anyone me and my partner’s batting average.
I use it to do things at the same time as other things. I can add something to the shopping list when I’m cooking or turn on the fan when I’m getting ready for bed without stopping what I’m doing to click buttons. I find that it’s really good for things that can’t easily be automated but you also can’t (or don’t want to) put on a physical button.
I actually use Alexa as I haven’t had time to investigate HA voice control but the principle is the same.
No Alexa here for more reasons than just being a voice assistant.
None of them would particularly useful to me even if I didn’t hate them, due to a lot of hearing loss. Real voices are hard enough and the ones from a speaker are worse. Anything from a PA system is unintelligible.
Most things we have automated with sensors, schedules and buttons. Not much else to be done that can’t be done with a few seconds of infrequent manual input, or shortcut software button on device.
I think I’m dead set on this. It’s nice to see others getting along, but I’ll never want it.
I guess they have watched a lot of Ironman and want to recreate Jarvis at some point
