Like @dnschnur, I’m curious to hear more about hardware support. And also like David, I would love to see this on Sonos. Sonos does support different voice assistants including support to run more than one simultaneously. While their developer portal doesn’t talk about how to create your own voice assistant, perhaps they would be responsive to an inquiry.
Looking forward to learning more!
Number one sentence in every language!!!
Yes dear!
This is very positive news. Some of you are looking at it from the point of view of Google Home or Alexa where your language support is, but in my little country where Google support was promised in 2018 and nothing has happened to date I will be very happy if Home Assistant goes in this direction and I will be able to at least partially control my devices by voice in my native language.
Would be a cool collaboration if we can use the ESP Muse Luxe speakers with microphone as the smart speaker.
I think I too would be hesitant to replace the numerous Google hubs and other voice capable devices around the home for something less pretty and less functional. I know of other automation systems that tried this and abandoned it, because it requires a lot of effort to maintain and run.
I doubt you’d ever be able to match the level of integration in smart speakers, smart tv’s, smartwatches etc. of Google, Siri and Alexa too. Why do you think you can succeed where those have doubts (I do not think they’d all go actually as far as abandoning it).
That’s kinda ironicly funny, I’ve worked the whole of 2022 to finally get away from any voice commands, to have plenty of sensors, with the sole purpose that I don’t have to command my house anymore but that it functions as autonomous as possibke, now you guys wanna focus on spoken commands 
Sounds like an interesting project anyway, good luck
Wouldn’t matter come in handy at this point ? Or will this actually test the matter intent and break it’s promise?
If ever I would have opted for something like this:
a solution which would allow me to talk to HA locally … and it auto creates an automation for me. Means not trying to locate the bug while writing yaml but simply talking to HA that it didn’t work as expected until it finally does work.
For the ones able to understand german language, there’s a far more amusing explanation about the capabilities.
This is interesting. Like many I have Alexa’s around the house but I’m getting fed up with the “by the way” suggestions. I contacted amazon to report it but they just went with the usual “have you tried a factory reset”. I was hoping Mycroft would be a good solution but it’s too expensive.
Worth noting though that I don’t use voice commands much. It’s much slower than just pressing a button. The main use case I have for this is while driving where I’m using Google Assistant instead. Unfortunately I recently asked it for the temperature on the thermostat and it started listing off every temperature senor linked to my smart heating.
I would be perfectly happy with a way to trigger a script from a specific phrase like “turn off away mode on the heating”. I think this is already possible for “turn off” phrases, just need to work it all out.
I looked into a local voice assistant before but i read everywhere that it’s really hard to get good microphones that are able to pickup your voice from the other side of the room.
Anyone has some good recommendations?
I think the current version of Rhasspy is already capable of that.
How do people feel about integrating an AI like ChatGPT into home Assistant when in the future it becomes (commercially) available to send API calls to. One could speak a prompt into the HA voice assistant with the request to ask chatGPT something, and the output from the AI would be converted back to speech by HA.
To me, that would really be frontline tech, would be the future.
Interested to hear the rest of the community on this topic.
Some integration will probably be made for ChatGPT, but this is out of the scope at the moment.
The goal is to make a local only voice assistant, so no outside resources needed for handling the voice commands.
I think you need a bit more detail. There’s no end of ‘good microphones’ Blue Snowball comes to mind. I feel like you have some requirements about style, placement and connectivity.
I found some info here: Hardware - Rhasspy
And that links to: Benchmarking Microphone Arrays: ReSpeaker, Conexant, MicroSemi AcuEdge, Matrix Creator, MiniDSP, PlayStation Eye | by Charles Rouchon | Snips Blog | Medium
There you can read that you cannot just use any microphone. I’m going to try finding an Playstation Eye to test with…
I tend to disagree.
A need for a dashboard is just showing that you need to interact with your system too much. To make it more smart it should have more automations than displays with dashboard interactions. If you need a temperature you should be able to get it anywhere in the house rather than approaching a display or unlocking a phone. In that sense voice might have brighter future and is not available in HA at all (other than 3rd party integrations).
Or just hacking it and preparing a custom firmware for existing hardware? Has anybody tried that already? Quickly searching for it I’ve found this page:
Edit: after browsing the article (and the second part) hacking it is utterly impossible… So only replacing the inside with different hardware could make the housing usable (or maybe also some parts e.g. the microphone?)
It’s useful and I agree with the idea directional mics are probably a poor choice, but I find it pretty inconclusive. The Blue Snowball I suggested is an array mic (of sorts) which can be set to be omnidirectonal, cardioid or and is advertised to be used in recording room scenarios with multiple speakers. (It’s also not cheap). I’d say, from reading that article, there are clearly a vast amount of untested microphones and they’ve also got a clear bias towards building a custom device using a dev board.
this would be great,
but still there is the microphone issue,
it would be great if I could use my phone for this.
so “Hé Rhasspy: turn on the lights”
Not sure if ios supports this, so then it would be “Hé Siri, open HA, Hé Rhassphy…” and that would be to long.
You can use your phone to turn on/off devices now if you install the Conversation integration. A microphone icon will appear in the top of the browser interface to Home Assistant.