Homey AI voice assistant coming in 2025?

I’m not sure how I feel about that video. Kind of great and creepy all at the same time :sweat_smile:

1 Like

Hopefully not. The Homey Pro isn’t capable of running any decent AI Model locally, and I wouldn’t want any of my IoT to connect to a public AI service. Maybe I could connect my Homey Pro to an AI model running on a local server somehow. But I would definitely not want them to build some public AI server feature natively in Homey.
AI is a hot buzzword now and companies are rushing to put anything out in order to not miss the boat, but tbh most companies really build crapy features that aren’t making anyones lives any better.

1 Like

I think the answer is somewhat clear as of 2 days ago, although I still don’t understand how this thing is going to work without haveing their own phones to put it on :thinking:

"In addition, LG Electronics is also stepping up its AI home business by applying Microsoft’s software to its mobile AI home hub “AI Agent” (project name Q9). After meeting with Microsoft CEO Satya Nadella on the 26th, LG Electronics CEO Cho Joo-wan said on his social networking service (SNS) that “we introduced LG Q9, powered by Azure OpenAI and Bing Search today. Through this partnership, we will be able to get closer to realizing a zero-labor home for AI.”

1 Like

In the most recent email from Homey there’s a link to a survey to “help shape the future”.

I found question #9 to be particularly interesting.. :grinning_face:

Where’s the €0.99/month option? Which is I think is still too expensive :zany_face:

5 Likes

Interesting to know how this project went..

2 Likes

Looks like the Project from the Kickstarter with the Homey (Not yet named Pro)

There is written a lot on this community (and the old forum)
Read their Blog/Official statement: here Our take on speech & music — Behind the Magic | Homey

TL;DR:

Simply put, we wanted a voice-controlled home with support for all smart devices. We underestimated the amount of work to solve speech, and are glad that this has been solved by other companies. From the moment we started developing Homey, we noticed users wanted multiroom speech recognition. This can now be realised with the products by companies such as Google and Amazon. That leaves the smart home to Homey, accompanied by an Amazon Echo or Google Home for those who want (multiroom) speech.
You can check out all the possibilities with speech integrations here.

1 Like

Indeed, @Dijker.
Also because Athom has moved from Rigtersbleek-Zandvoort to Oude Markt in Enschede long ago.
This is an old application…

Ah thanks for clearing up.

I was confused by the publication date ( feb 2025 )

Decent Local voice is one of the only things im missing currently. (Besides a few US vendors like Dreo (talks confirmed) Wyze (nothing as far as I know?)) And sure there are options but they are cumbersome to say the least.

For now Im still using my Google Nest (minis/hubs/hubs max) to voice control things. But that simply is not local (right?). If that could be local only for my queries please teach me :slight_smile:

I would love a local Homey first voice approach but not sure how much in the cards that is. The survey that was sent out shows they are thinking about it.

But that pricing … I get it’s endeavor but having it onboard would make Homey even so much more appealing to such a vast array of potential customers. Attaching a subscription to that, while for the Homey Pro the whole premise has been a one time purchase (ok besides backups if you need that). Making this a subscription would put a lot of people of,

Think of it like this;
How much more Homey devices would you sell thanks to that being a feature?
As apposed to how much of those subscriptons would you sell to current Homey owners AND people already interested in Homey (with them knowing it’s a return cost)?

For me personally; Homey should stay in the business of selling (and extending) their hardware.
Not become another company on the lets make subscription based software addons for our hardware.

Keep people you have motivated to be your brand ambassadors to others … nothing sells better than credible enduser word of mouth. (I think I have convinced 7 or so people to get into home automation with Homey, and that might seem small of a number, but how many subscriptions would you need to sell to make up for that?)

1 Like

The idea behind the subscription is that local voice, especially when “AI” is involved, will require too much processing power to be able to run locally on Homey. So like with Nests, Alexas and HomePods, it’ll have to be done somewhere in the cloud.

2 Likes

Agreed for sure Robert.

But that is key, for me local voice does not have to be “as good as“ AI can be.

Local Voice should be what the name says; local.
If we could use it for starting / ending flows, checking available stats/status that Homey already has/knows or can get from local integrations … that would be more than fine for most people.
That would be local. That still involves a lot r&d, testing, training etc for sure.

I don’t need it to do stuff like AI really can. I don’t need a x word amount essay or a comprehensive deduction of my device or insights over x time. I need it for local things, simply too be an easier or alternative way for me to get a result, that could have been gotten with human interaction through clicking in the app/a button/etc.

Just saying; if Homey thinks local first, voice should be approached local first as well.
Still would require quite the effort/processing power to accomplish, im not undercutting that. But that the entire challenge.

We simply don’t need it to become the next Nest/Echo/Homepod … because at that point … yay another one of those. … well whats the point even in that,

Hope the above makes sense :wink:

1 Like

The old Homey firmware has exactly that, and Athom removed it because it wasn’t great and took up too much (development) resources.

In the current “AI everything” climate, users expect that they can talk to their device conversationally, and I doubt very much that Athom is interested in implementing something like that themselves, which is why they are gauging interested in a paid (cloud) subscription for this.

1 Like

Upfront; Ok sorry for the rants above and here, but I come from a good place. I want Homey to think local with voice. Not just or not only jump on the cloud AI bandwagon.


Didn’t know the old firmware had something like that.

I get the conditioning that is needed for users to understand the difference between conversational and actionable approach to voice. Two edged sword for sure which ever way you flip it.

Sure I do understand cloud would be a way better fit for conversational voice use. But then why not even just be a connector / adapter / input source for the various available options that already exist (and a lot of people own). Being yet another (and even perhaps paid subscription) one would be so less more worth the (development) resources (at least to me). Just capture mic input and send it to a desired translation layer which then handles the handoff back to (for instance) Homey.

That in itself is nothing new, and what you can do without a subscription with a (for instance) Google Nest. Say something, Google understands, starts flow x on Homey. Tadaa, Voice command for Homey.

But NOT local, of course. So no internet no voice command.

So making you own well .. having a bit more control would be one I suppose but reinventing the wheel of speech recognition/algorithm for that just does not seem that logical to me. No offense but they are not going to beat Google/Amazon/Alexa in that front anytime soon. Im not saying they can’t but not out of the gates.

But fair they mentioned “AI voice assistant“ in the survey. If voice for Homey is going to be; ah yes let’s put it through our own (subscription paid) cloud AI service, well frankly im not even interested anymore in the potential voice solution.

What would be a main advantage of a Homey AI only cloud assistant? All there currently is, is cloud ai based voice assistants. What can we gain for on what we have now? Nothing substantial as far as I can tell.

Local voice is a limited scope, as it rightfully should be.
But Local voice is not easy to handle on hardware where it wasn’t a key point for (if it was… then correct me please)

Is it possible, yes, look at what Home Assistant is doing; Home Assistant Voice Preview Edition - Home Assistant
I would pay for a hardware add-on by Homey (like I have done with the Ethernet Adapter, with the Energy Dongle), which makes Homey Local Voice possible. But only if cloud was an option not required (kinda like HA is doing)

Either way … looking forward how this evolves and hope they do not choose the cloud AI only route.

Yes, that’s is a very good solution, as it gives the user the option to either have limited voice support, but fully local on limited hardware (similar to Homey), provided your language is supported, or much more elaborate voice support either requiring quite a beefy local machine, or using cloud services.

However, I can’t imagine that Athom is interested in the limited (speech-to-phrase) version of this setup, at this is basically what they already had, and cancelled. I think that if they implement speech support, it’ll be with a similar device with fully cloud-based processing. The typical Homey user doesn’t care about privacy, or being dependent on cloud services.

Agreed Robbert, and honestly I think they have made their decision as you also stated.

But still;

Wishful me; Well I would love if they offered both. Options are always the way to go.

Realistic me; Well they are a business, they will go the route which offers the most upside for them. But that all comes down to selling more Homey’s on the long road, what drives that more looking at this, well I can guess but I have no control over that.

Pessimistic me; Well might have to look into HA hardware then and see how good that would work when combined. Don’t like mixing the hardware bits from both Homey and HA all that much but hey if that is the route that is needed, well , we will see how that works out; While we are on that matter, has ANYONE combined both of them yet and can you give us some insight on if that works well?

Proactive me; Well there must be a way to still get local on the table or at least get some official feedback on their stance/reasoning. Lets debate, lets point out stuff, lets at least keep the enduser side rolling.

As for privacy; meh that ship sailed for me and a lot of people way back. If you are on the internet or connected to the internet, even how hard you try. Privacy is such an illusion to start with.