Home Assistant made their own Alexa!!

[@networkchuck_v2]

Home Assistant is finally taking on Alexa. They sent me early access to their new Home Assistant Voice. It’s like Alexa….but way smarter. This thing uses AI to control your ENTIRE smart home, and in this video, I’m unboxing it, testing out its features, and sharing my thoughts. Can Home Assistant Voice really replace Alexa? Let’s find out! 🔥🔥Join the NetworkChuck Academy!: ntck.co/NCAcademy 🛠 Explore Home Assistant: www.home-assistant.io/ **NOT Sponsored

[@Path.V]

Two things, update your stuff because you can set a pipeline to do local stuff first then fallback to ai if needed, think turn on lights vs when did the matrix come out. It's in the pipeline settings. Second thing, the Wyoming satellite can be set up as a media player as well! Not too hard to run from music assistant! Good luck!

[@urmastertech]

Oh dang how do you setup the satellite for the media player? That would be awesome for mine plus seeing if I can use a better quality speaker or just plug it into a speaker.

[@FlorianGT396]

​@@urmastertechMedia Playback is not possible i think. Thats why i stick to my pi zero.

[@urmastertech]

@@FlorianGT396 ah ok that makes sense. I suppose it's just a raspberry pi so I could just install something on it to allow that. I only set it up like once so I haven't used it much since

[@FlorianGT396]

@@urmastertech i was wrong. Media playback from for example musicAssistant is possible. And the hardware is a esp32. But If you like raspi base more there are also possible ways.

[@urmastertech]

@@FlorianGT396 I would actually consider the esp32 way too. I have one laying around, but seems like I would have to get a new one that has the micro voice thing

[@CraigTinson2015]

I ordered mine the minute it was released.. looking forward to playing with it.. see how it compares to my s3-box... also, in the pipline.. you might want to click the new "Prefer handling commands locally" .. it might help with using Terry.. it'll try to figure it out first.. before passing it to the LLM...

[@urmastertech]

This is awesome, I hope they come out with a v2/production release of this. I am tempted to pick it up though to see how it compares. $60 is a little higher than I thought, but still not a bad price.

[@02semiata]

Probably going to be software related updates for awhile. Don’t expect a v2 hardware release

[@JoeyG1973]

So what's a Grove port again?

[@Obtuse94]

Expansion port, can add sensors, a display, etc

[@hellsdark]

Very cool. I will probably buy one as soon as my language is supported

[@rasbe6863]

You have to provide context to what light to be turned off. If you're trying to turn off a light in a different area then the area you assigned. So you assigned it office. When you say turn off the light, it's going to turn off the light associated with the office, not the one in the kitchen

[@FreedomToRoam86]

I want this, way more than Alexa. Being able to give troubleshooting info when it can’t do something would be dope! And would be even better if could do HAL voice, and open the pod bay door…. 😂

[@domramsey]

I'm pretty sure that 50% of your issues were caused by not updating Home Assistant beforehand (and using the newer local processing for HA commands functionality), and the other 50% were down to your own random LLM. Enjoyed the video, but I do think if you're going to review something, it's helpful to make sure the things it depends on are in good working order first. Otherwise, it's not very fair to the product.

[@superbanane4768]

Does anyone know if the wake word detection is done on device?

[@Lamp512]

the wake word is 100% local on the device

[@noelflores8989]

Update your haos to get the new voice functionality. It can try haos first before going to your ai pipeline

[@CraigTinson2015]

and the grove port is for adding any of the grove sensors from the Seeed Studio line-up.. among a few others..

[@NathanBudd]

"this is my giddy face" 😆

[@CPPRODUCTIONS1001]

So a lot of the issues seem to be with custom ollama stuff? Hopefully thats fixed but if it works well with the core HA assistant, that seems good

[@SamBrownhousehold]

it would be interesting to see if you had a few of them would they all trigger or would they figure it out like alexa does? we have the issue of a small flat so all the echos in every room can hear you no matter where you are.

[@nabaroy5356]

I am so pump when you talk about Home Assistant

[@nabaroy5356]

This speaker need to get upgraded

[@AcheForWake]

What are you going to do with this box? 1:00 And how does the device look? 1:03

[@malvo_b]

I would like to know the cost when they release to the public. Kincoy makes a really good voice assistant device that’s loud and it’s only like 39$

[@ffurtech]

its 59$ right now on HA's website

[@thestreamreader]

Does it work better than this device??

[@malvo_b]

@ I personally haven’t purchased it yet . But from what I seen in a KZbin video it does work better and is bigger .

[@malvo_b]

@ about the same as the atomic echo he initially showed which isn’t bad .

[@That1FamousWRX]

Jesus Christ, It’s Jason Bourne!!

[@DragonXVI]

It’s a mountain of potential. Hopefully the bugs get worked out in the release version. Did they mention anything about a price point?

[@sean640]

What would your limit be? Not over $40?

[@CraigTinson2015]

$59

[@ericbaker8807]

I think it's subject to retailer pricing... But the Home Assistant / Nabu Casa folks have a recommended MSRP of like $60. I just ordered one from AmeriDroid for like $63. I can't wait to play around with it.😊

[@alexitanguay]

I wouldn’t pay more than 50$ CAD.

[@CraigTinson2015]

techncally I paid "48" UKP.. sound better? :P

[@itsp0c0vt]

I WANT THIS. PLZ NABU!! I Love Home Assistant with a capitol L for my continued reliance of Alexa to provide me voice assistant integration. Thi would be an amazing product and would recommend it to anyone in New Zealand who wants to get into privatized smart home setups

[@MisplacedAmerican]

What do I think? I think I NEED one!

[@CampRusso]

That is cool! I'm guessing the 3.5mm you could connect a big powered speaker? 🤔

[@DashDapper]

Will definitely be getting one to play around with. The whole "this talks to Jeff Bezos" sold me. Would like to know about it's functionality with Spotify, KZbin music, SoundCloud etc

[@itsp0c0vt]

I reckon if those services can be added to Home Assistant as a media source integration, then the HA Voice should be able to play it.

[@leekirrage]

I have Spotify integrated in to Music Assistant on Home Assistant, easy to do so this will work.

[@Ali_9402]

It's just an start to build my own assistant, when you can build it yourself I wouldn't lose the joy

[@Tokamak_9]

Alexa, you are redundand. .... crying echoes in the speakers.

[@Seedlinux]

I am ready to pay, I'll take one

[@WreckDiver99]

Giggidy Giggidy Giggidy...That's chuck....Hmmmmm.... I am looking at creating something like an info dashboard with a spare 40" TV we have (from my mother-in-laws, no we didn't just buy an extra). I really like what Thomas Sanladerer (Made with Layers YT Channel) did, but I want something that does even more or is just more "flexible". Wondering if Home Assistant can do some of this.

[@Cyber_vir]

I sold man. Just ordered one

[@SorceressEmmaly]

This is your giddy face.

[@XBitX]

Hey Chuck! I noticed that you have pending updates for Home Assistant Core and OS in this video. When you were working with the timer and it not working with Ollama, specifically when you were changing the settings on the Conversation Agent, I didn't see one of the new options they added in a new update that _may_ have helped a bit here. The new option is directly under the "Conversation Agent" selection and is called "Prefer handling commands locally" with a toggle. What happens is the Home Assistant default local conversation agent is ALWAYS the first stop for a command no matter what agent is selected when the option is enabled. If the command is not understood it will then forward it to the agent you selected. This would have probably helped with the timer since it's a built in function, and then things like "When was the first Matrix movie released" would have went to your Terry Ollama agent. It's not perfect, as you've noticed that the commands are very specific wording right now, but they add new ways of saying things all the time. So while it's not an LLM handling everything for you, it does cut down on some of the hiccups and confusion that LLMs currently do when hallucinating or otherwise. Otherwise, good video and I'm hyped for the new hardware!

[@SweetHome-w8c]

Make a video about AI in OSINT

[@donaldduck7268]

Can you improve the voices somehow

[@Josh-vu4sb]

Well to ebay the raspi goes XD

[@memthemems1026]

Alllee chsaaa. My vooooiisse iss breakknnn

[@bobiscool5466]

Epic

[@yobe7906]

It’s a bit disingenuous that you cut between the time you ask a question and response. First I thought you were running a monster build.

[@DeepapgbTk]

Thanks for sharing!❤ I need some advice: My OKX wallet holds some USDT, and I have the 12 words: check broom dinosaur stool harsh hurdle morning solve barely umbrella crumble naive. What's the best way to send them to Binance?🤔

[@smurththepocket2839]

Wrong video buddy

[@superangrybrit]

Promising. It needs to be a smart plug format tho. Very nice beta product. 👍

[@SpaceDoodle2008]

Your second channel should have been called NetworkChuck[1]

[@SpiritsAtWar]

Can I drop ship this product for you?

[@Jules-hn6un]

Utterly underwhelming product. A poor replica of Amazon Echo but just with the “it runs locally” for all the paranoid schizophrenics

[@wizkidweb]

It's not paranoid. The reason why you can often get an Amazon Echo for $20 is because Amazon sells them at a loss, because they know they can make far more revenue from the metrics gathered by listening to your every word. They also sell this data to other companies, who can then use it for targeting campaigns. Some people care more about privacy, and this is incredibly impressive given that it doesn't require a data center or internet to work.

[@Jules-hn6un]

@ “listening to every word” but yeah, not paranoid 😂

[@wizkidweb]

@Jules-hn6un If you know how they work, that's exactly what they do. It only transmits when you say the wake word, but both the echo and the Home Assistant voice are always listening. The difference here is that the latter is totally open source, while we have to trust Amazon that the former won't do anything nefarious.

[@anwar.shamim]

not interesting device

[@spenchurch]

i disagree, anwar shamim

[@anwar.shamim]

@@spenchurch I don't see interesting once you have Alexa echo or Something else

[@sean640]

​@@anwar.shamim Google and Amazon make money from storing servers with data they collect of your voice info that's why

[@_Dreadd]

@@anwar.shamimthe point is that your voice assistant is hosted locally through your homelab, not by amazon. Its more customizable for your household since it can pull from private resources on your homelab. also im sure voice assistant generally has plenty more features than amazon echo.

[@jeffreyeaton5214]

@@anwar.shamim The biggest benefit is you don't have a trillion dollar corporation spying on your and selling the data it's harvesting from spying on you. No, it might not be as fleshed out as an Alexa, Google, or Siri device, but at least you can have peace at mind knowing your life isn't being invaded and info being sold. Pros and cons

[@oildiggerlwd]

It looks like some of the other reviewers I’ve seen today didn’t have an LLM hooked up to it either locally or through an API. This review was far more positive of it I think because it had a LLM hooked up Surprised @technithusiast didn’t have Ke hooked up.

[@Leuviah1337]

Epic