It's here! Local voice control with WAKE WORD in Home Assistant! Tell me what you think. All the links mentioned can be found in the description of this video. NOTE: This tutorial also works for the Atom Echo!
I was sitting on my hands waiting for this very feature for HASS. I know its in early development but I want to say thank you so, so much to the HASS dev's!
This functionality is great! Im planning on getting an RV for permanent living situation and this would drastically make it easier to have all the same smart home comforts
Thank you so much for your kind words! I'm glad you find the functionality great. Best of luck with your RV journey and may you have all the comforts of a smart home on the road!
There is an easier way of creating a "custom wakeword"...by installing "snowboy" in the add-ons...its quicker and you say the new wake word 3 times and learnt your voice...just select it in assistants menus
Nice. The demo software that comes installed on the device does wake word detection on the device, so there is clearly the opportunity for the ESPHome solution to take us there as well.
Just watch the Espressif videos on this box. Wow there are so many features built straight into the unit. I can't wait to see some of the future projects next year.
@@SmartHomeJunkie Unfortunately, I don't own it - I was patiently waiting for the wake word and now that it's here, I have to wait for the HW :) Well, I deserve it...
@@SmartHomeJunkie I am not very consumery-type, so I am ok with waiting a bit - at least the SW will get even better by the time I get it ;) But it's great to see people like you doing the walk to make sure lazy people like me have a clear path towards the target ;)
Great video, nice to see the out takes at the end. It's still early days for this product, it will be interesting to see what improvements can be made by using the second microphone.
so happy wake words finally work! I hope the S3-Box will be available again soon... Do you know if there is a way to build something similar in a dyi fashion using a ESP32, mic, and a small screen?
Yay, the wake words are finally working! 🎉 As for the S3-Box, I'm not sure when it'll be available again. Not sure how to build something like this yourself though.
I appreciate your video. When running locally or with Home Assistant Cloud, it can be much faster than Alexa and Google Home for commands. But to be honest, the ESP Home track is a side track. This would gain much more momentum if you could use wake words with dashboards. We all have dashboards already. For instance, have some old Windows 10 tablets (Toshiba Click Mini) where I run web apps for Home Assistant and Alexa side by side. One with wake words, one without. I'd love for it to evolve further. And, I guess it might.
Hey there! Thanks a bunch for watching and appreciating the video. It's awesome that you find Home Assistant to be faster than Alexa and Google Home. I totally get what you mean about the ESP Home track being a side track, but hey, sometimes it's fun to explore those side roads, right? And trust me, I'm all ears when it comes to using wake words with dashboards. Let's see where this evolution takes us!
@@SmartHomeJunkie Yes! And what would Home Assistant be without all these side roads? As a matter of fact, I am thinking about testing out the Permobil integration. I don't expect that has the largest installation base either.
Honestly, I prefer using the case swapped Google Nest mini. There's a custom PCB you can order to put in the it's case and use it for this. Most Mics I would use for this get no benefit from having a screen. Maybe for my desk, but I'd want to be able to give voice commands in multiple rooms.
Most of us will not ditch anything. Google home and Alexa are very simple to set up. HA is extremely difficult to set up, I have been messing for over a year and am still getting nowhere.
This is great and will certainly be interesting to many Smart Home enthusiasts. However, I don't see this replacing Assistant or Alexa, because they do more than just control smart home entities. Playing media answering questions, controlling TVs and other entertainment Devices and of course setting timers, reminders and calendar entries are features that are unlikely to all fit into Home Assistant, so I see this as an addition, not a replacement. In a perfect world, Google or Alexa Speakers would be fully integrated into Home Assistant and would allow us the best of both worlds...intuitive control of our Smart Home and all the additional stuff these "smart" speakers do.
Theoretically, yes. However, both Alexa and Google Assistant have deteriorated badly. Basically, the only thing they are used for in my house is for the weather and a voice front end for HA. And that isn't uncommon.
Do you think it has degraded ? I often think the same but I wonder if is just that it has not improved and I expect it todo more. I was amazed when chatgpt came out and I was asking it programming tasks but now it frustrates me more than anything.@@Banzai51
Hm, combining these sensors with Agara presence sensors opens up great possibilities because HA will know where you are when you say a command. Will not be cheap though.
Thanks for sharing your thoughts! Totally agree, it would be awesome to have a smart speaker that matches the sound quality of the Echo or Apple HomePod. Let's keep our fingers crossed for some cool new options!
why we cant connect an BT Speaker with the S3-Box? The Box self has BT integrated? Dont understand why HA didnt using this with the spezial firmware. I connected my Walldashboard with BT with my Alexa Speaker. And the Assistant from the Dashboard has the Jarvis Wakeword Too. Thats my solution for now... but more likely i would use the new S3-Box or sth like that.... very unlikely that such a nice product has such a terrible speaker @@SmartHomeJunkie
Hello. Thank you for all the videos. I am having difficulties adding music to my home assistant. I'm looking for something that doesn;t have a cloud but can paly a radio for exampe
I just can't get the S3Box3 to work I have no idea what I've done wrong I can load the firmware onto it but then it wont boot up - the screen is blank I am fed up stuffying around with it frankly I just wish NC came up with a hardware voice device that darn well worked out the box rant over!
I understand why HA added this... but I honestly rarely speak to my Alexa now. My smart home runs without needing to speak to it, which is the goal right?
Please create a guide on how to integrate AI into the assistant, as I often find my current language assistant unreliable and requiring precise phrases, which is not an optimal solution. I have read that it is possible somewhere, but I have never successfully implemented it. Therefore, Google Assistant remains a more suitable option for me, as it understands a broader range of phrases.
You know what? I Ajay created a video about that here: How to set up Voice Assistants / What's New in Home Assistant May 2023.5? Voice, Voice, and Voice! ruclips.net/video/nh3jnPGAtMI/видео.html
So I bought one of these ESP32-s3-box3 and got it all working. Then 6 hours later I smelled a terrible smell and the esp32-s3-box3 was dead. Zero support to get a replacement. We need a real product, not crappy hardware that dies in hours.
I'd love to get rid of my google devices but the physical equipment to replace voice and sound is so poor it's not feasible. What equipment can I really place in each room to replace google?
Ok, so now it's time to use some USBc speaker/mic device as "speaker" and make a use of ChatGPT-"like" assistant to ask more complex Questions (and forget limited Google) That kind of manual will be grate!
@@SmartHomeJunkie hi, yes, i know this video ;) but you'r talking about AtomEcho same as ESPdisplay, and i'm thinking about just simple USB mic/speaker (like Anker conference speaker) always connected to main device and listening ;) but till now i cant find any info how to configure this out i inside HA (only way for now i found is Rhasspy)
Should work on any esp32 device, no? Thinking about trying it on an M5Stack. They’re still in stock here 😅 And if it doesn’t I‘ll still have a decent little device to play around with…
I would love to be able to reflash my echo devices with something like this. And maybe, some AI and services integrations via HA would finally put Alexa in the graveyard
When I try and flash the Atom Echo from the ESP Site it just gets stuck scanning for networks. Any ideas how to get around this. I can add it to wifi in ESPHome directly, but I assume it won't have the voice software on it then.
What is the advantage of a one single local only box compared to the wide spread commercial ones? Correct me if I am wrong, but all you can do with that box is to control your home devices and just from one location where that box is placed.. Is that what you call an assistant?
This video is meant to show what can be done by something still in development as of now, not as an end product to compete with Google and Amazon. It is still useful to many as of now regardless, especially to those of us who don't like to buy products that make money off of spying on you.
How do you find the limited number of words for offline speech recognition? From memory it is 200, which sounds like a lot but I bet they get used up really quickly and how do you know easily if it is in online or offline mode so you can simplify your diction?
thanks for the video...I have watched your channel for a couple of years...I really enjoy, but I am tired of buying hardware for my smart home...I have 6 different Google hubs and dots in my house...and I go thru Nabu Casa now...I am sorry to be blunt, but why would I buy this?
Ah, the age-old question of how far it can hear you! Let me tell you, this device has ears like a bat... well, not quite, but it can hear you loud and clear from a distance of up to 4 meters. So, you won't have to worry about shouting to get its attention! Well, that's the range that I tested anyway.
Can this be tied into stuff like search as well? I’m getting so tired of Google randomly activating at different times and not answering on one home mini and then the other answers and so on. I’d love to be able to use something reliable for stuff like automations and commands but also the convenience of being able to do a quick search and have the results read to me or show a recipe on a display or something. Like would I be an to set this up to be a chatgpt powered assistant?
I'm on ESP Home 2023.12.9 and getting the "The connection dropped immediately after encrypted hello; Try enabling encryption on the device or turning off encryption on the client" error message. I've flashed my device from the proyects page and proceeded to add it to Esp Home, but I'm not able to compile a binary from my local ESP Home in order to flash the device with an encryption key.
Figured it out. Even though I had trusted expensive cable, turns out it don't have all Type-C pins. It worked with a cable from the original BOX-3 box :D UPD: It worked once... very unstable, can't repeat it and constantly have same issue :( Also, sometimes it boots up after flash, sometimes stuck in black boot mode...
Hello! I would like to ask what kind of hardware the voice commands run on locally, so that the response time is so fast? How powerful is your computer?
Hi, thanks for this interesting video, but I'm encountering an extra step. I'm using an atom echo and he is asking for an encryption key at approx 5:42. He shows the following, "The ESPHome device Atom Echo enabled transport encryption or changed the encryption key. Please enter the updated key. You can find it in the ESPHome Dashboard or in your device configuration." Can't connect to ESP. Please make sure your YAML file contains an 'api:' line. Could you please point me to a solution
I've got a very specific question. If you setup an ESP32-S3-BOX in a specific area (for example my "Office"), and you also assign it to this area in Home Assistant. If you then use the command "Turn on the lights", does it only turn on the lights in the assigned area (in our example the "Office" lights)?
Do you know if there's anything that needs to be done in order to dim the lights? My lights turn on and off fine but when i request to dim to a percentage it doesn't understand. I'm using an Atom echo. Thanks
I got the atom echo with a view to getting rid of google, but the quality of these things are not even in the same ballpark as google, it will take years of development to even consider comparing them.
Hey, thanks for sharing your experience! It's true that the Atom Echo has some ground to cover before it can compete with Google. But hey, every journey starts with a single step, right? Let's hope for some exciting updates in the future!
Yes, it's a fun/torturous journey haha. from installing ha to atom echo to frigate, nothing sails plainly, but what a sense of achievement. I have my own netflix now, if only I could have my own spotify. Is there anything that can, say....'hey junkie play music', I can't find anything that will just play my server music via voice control :(.....thanks for replying, you make great videos...@@SmartHomeJunkie
Awesome. Just wish HA was easier to use on Windows. I don't want to use a Docker container or VM, or a Linux-on-Windows system, and I don't want to fight with compiling the shit myself. Native support please!
@@TheBroz migrate? No. But the platform isn't doing anything unique to Linux. You realize projects can be coded for multiple platforms in a single code base, don't you? Since there's no/very little graphics related things, I don't imagine there would even be all that many OS checks needed.
@SmartHomeJunkie do you know if it's possible to assign which audio device for the assist pipeline tts responses the audio plays through? I am running a mimic3 server on a different machine with marytts configured to talk to it and I have several audio speakers including a hass agent and sonos speakers I would love to use to play the audio through (instead of the device listening for the voice commands)
It's here! Local voice control with WAKE WORD in Home Assistant! Tell me what you think. All the links mentioned can be found in the description of this video.
NOTE: This tutorial also works for the Atom Echo!
I was sitting on my hands waiting for this very feature for HASS. I know its in early development but I want to say thank you so, so much to the HASS dev's!
I'm not really interested in voice recognition, but I love that you also took the time to write this down in your blog. Have my like and subscription!
I'm glad you appreciated the effort I put into writing it down. Thank you for the like and subscription!
This functionality is great! Im planning on getting an RV for permanent living situation and this would drastically make it easier to have all the same smart home comforts
Thank you so much for your kind words! I'm glad you find the functionality great. Best of luck with your RV journey and may you have all the comforts of a smart home on the road!
There is an easier way of creating a "custom wakeword"...by installing "snowboy" in the add-ons...its quicker and you say the new wake word 3 times and learnt your voice...just select it in assistants menus
Wow, I had no idea! Thanks for the heads up, gonna try it out ASAP!
Nice. The demo software that comes installed on the device does wake word detection on the device, so there is clearly the opportunity for the ESPHome solution to take us there as well.
Thank you for sharing your thoughts! Let's see how far the ESPHome solution can take us!
The Raspberry Pi versions for HA assist satellites can do this but I doubt we will get it the ESP since they are not overly powerful
Just watch the Espressif videos on this box. Wow there are so many features built straight into the unit.
I can't wait to see some of the future projects next year.
I know, right? It's like this unit has more features than a Swiss Army knife! Can't wait to show you new things for next year!
I'll test it once the device is available again :) And I will buy you coffee afterwards. Thanks!
The tutorial also works for the Athom Echo. Maybe you already own one of them? And thanks for the coffee!!!
@@SmartHomeJunkie Unfortunately, I don't own it - I was patiently waiting for the wake word and now that it's here, I have to wait for the HW :) Well, I deserve it...
@@rklauco This is always the thing... Some new device is on the market and then it's sold out immediately... Pffff
@@SmartHomeJunkie I am not very consumery-type, so I am ok with waiting a bit - at least the SW will get even better by the time I get it ;) But it's great to see people like you doing the walk to make sure lazy people like me have a clear path towards the target ;)
@@rklauco Hahaha! I'm glad I can demystify things for you.
Great video, nice to see the out takes at the end. It's still early days for this product, it will be interesting to see what improvements can be made by using the second microphone.
Wait what? Is there a second microphone????
Apparently so, isn't that what the two holes at the top of the screen are for?
so happy wake words finally work!
I hope the S3-Box will be available again soon... Do you know if there is a way to build something similar in a dyi fashion using a ESP32, mic, and a small screen?
Yay, the wake words are finally working! 🎉 As for the S3-Box, I'm not sure when it'll be available again. Not sure how to build something like this yourself though.
Digikey is saying early February for new stock.
You can flash the ESPHome voice assistant firmware on any ESP32 yourself. For a good mic, lookup the INMP441
I appreciate your video. When running locally or with Home Assistant Cloud, it can be much faster than Alexa and Google Home for commands. But to be honest, the ESP Home track is a side track. This would gain much more momentum if you could use wake words with dashboards. We all have dashboards already. For instance, have some old Windows 10 tablets (Toshiba Click Mini) where I run web apps for Home Assistant and Alexa side by side. One with wake words, one without. I'd love for it to evolve further. And, I guess it might.
Hey there! Thanks a bunch for watching and appreciating the video. It's awesome that you find Home Assistant to be faster than Alexa and Google Home. I totally get what you mean about the ESP Home track being a side track, but hey, sometimes it's fun to explore those side roads, right? And trust me, I'm all ears when it comes to using wake words with dashboards. Let's see where this evolution takes us!
@@SmartHomeJunkie Yes! And what would Home Assistant be without all these side roads? As a matter of fact, I am thinking about testing out the Permobil integration. I don't expect that has the largest installation base either.
Amazing tutorial!! There is a way to use this like actionable notification ( speakers send a message to take an action and )
Honestly, I prefer using the case swapped Google Nest mini. There's a custom PCB you can order to put in the it's case and use it for this. Most Mics I would use for this get no benefit from having a screen. Maybe for my desk, but I'd want to be able to give voice commands in multiple rooms.
Swapping the case and using a custom PCB sounds like a clever way to make the most out of your Google Nest mini.
This sounds very interesting! Do you have a link?
@@richardharnwell3331 Yes, though youtube usually removed comments with links. Just search "Onju voice google nest mini". It's the PCB Way link.
@@richardharnwell3331 Here you go: www.pcbway.com/project/shareproject/Onju_Voice_d33625a1.html
Isn't the Onju project abandoned?
I like your videos. I was excited to see this. Much less excited when I tried to purchase a ESP32-S3-BOX-3. Doesn't exist for months... :(
Thanks for another great video. Is this able to be done with another device ? Not available in SA, thanks
The only other device to use is the Atom Echo afaik.
Most of us will not ditch anything. Google home and Alexa are very simple to set up. HA is extremely difficult to set up, I have been messing for over a year and am still getting nowhere.
Also true
This is great and will certainly be interesting to many Smart Home enthusiasts. However, I don't see this replacing Assistant or Alexa, because they do more than just control smart home entities. Playing media answering questions, controlling TVs and other entertainment Devices and of course setting timers, reminders and calendar entries are features that are unlikely to all fit into Home Assistant, so I see this as an addition, not a replacement.
In a perfect world, Google or Alexa Speakers would be fully integrated into Home Assistant and would allow us the best of both worlds...intuitive control of our Smart Home and all the additional stuff these "smart" speakers do.
Syncing music as well.
Theoretically, yes. However, both Alexa and Google Assistant have deteriorated badly. Basically, the only thing they are used for in my house is for the weather and a voice front end for HA. And that isn't uncommon.
Do you think it has degraded ? I often think the same but I wonder if is just that it has not improved and I expect it todo more. I was amazed when chatgpt came out and I was asking it programming tasks but now it frustrates me more than anything.@@Banzai51
Hm, combining these sensors with Agara presence sensors opens up great possibilities because HA will know where you are when you say a command. Will not be cheap though.
The box s3 mentioned in the video has an optional expansion module with a presence sensor
Hi, thank you for your video! Can I use my four Amazon Echo with this method?
Nice, now there needs to be also a smart speaker available that actually sounds as good as an Echo or Apple HomePod
Thanks for sharing your thoughts! Totally agree, it would be awesome to have a smart speaker that matches the sound quality of the Echo or Apple HomePod. Let's keep our fingers crossed for some cool new options!
why we cant connect an BT Speaker with the S3-Box? The Box self has BT integrated? Dont understand why HA didnt using this with the spezial firmware. I connected my Walldashboard with BT with my Alexa Speaker. And the Assistant from the Dashboard has the Jarvis Wakeword Too. Thats my solution for now... but more likely i would use the new S3-Box or sth like that.... very unlikely that such a nice product has such a terrible speaker @@SmartHomeJunkie
How do timers work? It’s the one thing I still need Alexa for. Setting a timer when cooking
Hello. Thank you for all the videos. I am having difficulties adding music to my home assistant. I'm looking for something that doesn;t have a cloud but can paly a radio for exampe
I just can't get the S3Box3 to work I have no idea what I've done wrong I can load the firmware onto it but then it wont boot up - the screen is blank I am fed up stuffying around with it frankly I just wish NC came up with a hardware voice device that darn well worked out the box rant over!
Nabucasa Cloud - You make your self dependend again, but I want to say, to run the voice engine your self is quite recource hungry.
unfortunatly my costum wake word doesn't show up in the list somehow.. did check the spelling of the folder three times
I understand why HA added this... but I honestly rarely speak to my Alexa now. My smart home runs without needing to speak to it, which is the goal right?
Please create a guide on how to integrate AI into the assistant, as I often find my current language assistant unreliable and requiring precise phrases, which is not an optimal solution. I have read that it is possible somewhere, but I have never successfully implemented it. Therefore, Google Assistant remains a more suitable option for me, as it understands a broader range of phrases.
You know what? I Ajay created a video about that here:
How to set up Voice Assistants / What's New in Home Assistant May 2023.5? Voice, Voice, and Voice!
ruclips.net/video/nh3jnPGAtMI/видео.html
I had my s3 arrive today. Installed it by following your steps, but it is not listening to any wake words. Did you encounter this problem? Thanks
So I bought one of these ESP32-s3-box3 and got it all working. Then 6 hours later I smelled a terrible smell and the esp32-s3-box3 was dead. Zero support to get a replacement.
We need a real product, not crappy hardware that dies in hours.
can different users in the home have their own wake words, or can you only set one at a time per device?
I'd love to get rid of my google devices but the physical equipment to replace voice and sound is so poor it's not feasible. What equipment can I really place in each room to replace google?
Hello
After erasing all the data it’s not popping ip for wifi password so I was unable to connect wifi to my esp32-s3-3. I need help
Ok, so now it's time to use some USBc speaker/mic device as "speaker" and make a use of ChatGPT-"like" assistant to ask more complex Questions (and forget limited Google) That kind of manual will be grate!
See the video 2023.5 that I refer to in this video. I explain in that video how to do that.
@@SmartHomeJunkie hi, yes, i know this video ;) but you'r talking about AtomEcho same as ESPdisplay, and i'm thinking about just simple USB mic/speaker (like Anker conference speaker) always connected to main device and listening ;) but till now i cant find any info how to configure this out i inside HA (only way for now i found is Rhasspy)
Should work on any esp32 device, no? Thinking about trying it on an M5Stack. They’re still in stock here 😅
And if it doesn’t I‘ll still have a decent little device to play around with…
It works on the Atom Echo.
I would love to be able to reflash my echo devices with something like this. And maybe, some AI and services integrations via HA would finally put Alexa in the graveyard
When I try and flash the Atom Echo from the ESP Site it just gets stuck scanning for networks. Any ideas how to get around this. I can add it to wifi in ESPHome directly, but I assume it won't have the voice software on it then.
Can it be deployed on a Mycroft Mark 2? I know a ton of people will be interested since most are bricks now!
What is the advantage of a one single local only box compared to the wide spread commercial ones? Correct me if I am wrong, but all you can do with that box is to control your home devices and just from one location where that box is placed.. Is that what you call an assistant?
This video is meant to show what can be done by something still in development as of now, not as an end product to compete with Google and Amazon.
It is still useful to many as of now regardless, especially to those of us who don't like to buy products that make money off of spying on you.
it sounds like the box-3 is responding with just the last word that you say. is that something that can be turned off?
No, it's responding with sentences like: Turned Off and Turned On. There is no option as far as I know to turn it off.
@@SmartHomeJunkie ah okay! it wasn't easy to hear but happy if it's answering in full responses :)
How do you find the limited number of words for offline speech recognition? From memory it is 200, which sounds like a lot but I bet they get used up really quickly and how do you know easily if it is in online or offline mode so you can simplify your diction?
Thank you for bringing up this interesting question! I must say that I do not have any experience with it yet.
Where do you have your info from? It uses whisper in the background which has no particular limit.
thanks for the video...I have watched your channel for a couple of years...I really enjoy, but I am tired of buying hardware for my smart home...I have 6 different Google hubs and dots in my house...and I go thru Nabu Casa now...I am sorry to be blunt, but why would I buy this?
Uhmmmmm.... Because it's local and faster that Google Assistant?
Excuse me, but it work stand alone or must to be on even one PC?
When language is PT (Portuguese) the accuracy of understanding commands is not very good 😢
there is a solution for conversatoin monitoring like video surveillance that store the data in a database.
How far can it hear you from ? I've seen this done so many times with esp32 but if it will only clearly hear you from 1 meter, than it is useless
Ah, the age-old question of how far it can hear you! Let me tell you, this device has ears like a bat... well, not quite, but it can hear you loud and clear from a distance of up to 4 meters. So, you won't have to worry about shouting to get its attention! Well, that's the range that I tested anyway.
@@SmartHomeJunkie That's very nice. Will gladly try it then. Thank you very much
Can this be tied into stuff like search as well? I’m getting so tired of Google randomly activating at different times and not answering on one home mini and then the other answers and so on. I’d love to be able to use something reliable for stuff like automations and commands but also the convenience of being able to do a quick search and have the results read to me or show a recipe on a display or something.
Like would I be an to set this up to be a chatgpt powered assistant?
If you follow the link to the 2023.5 release video on the description, you can find the tutorial on how to set this up using openAI. 👍🏻
@@SmartHomeJunkie awesome and thank you so much! Have you tried this yourself yet or do you have any thoughts about it so far? Any tips when doing so?
@@muchachogrande2019 I show in that video how it works.
@@SmartHomeJunkie I understand, I’m curious about your personal experiences and thoughts on it with every day usage.
It sounds much like what I have noticed where the start of the voice reply gets clipped.
I'm tempted to disable the voice reply
I understand what you mean. A bit of feedback would be nice though. Maybe something like a simple beep?
For playing around it is nice, but I dont think it is very useful in day to day use. Its not a replacement for Google assistant.
Thank you for sharing your thoughts!
With olama or GPT integrations they are infinitely more competent then Alexa or Google. Without question
I'm on ESP Home 2023.12.9 and getting the "The connection dropped immediately after encrypted hello; Try enabling encryption on the device or turning off encryption on the client" error message. I've flashed my device from the proyects page and proceeded to add it to Esp Home, but I'm not able to compile a binary from my local ESP Home in order to flash the device with an encryption key.
That is a really old version...
After flushing Voice Assistant, for some reason it is not asking for wifi password... and any attempt to configure one through ESPHome page fails :(
Figured it out. Even though I had trusted expensive cable, turns out it don't have all Type-C pins. It worked with a cable from the original BOX-3 box :D
UPD: It worked once... very unstable, can't repeat it and constantly have same issue :( Also, sometimes it boots up after flash, sometimes stuck in black boot mode...
Hello! I would like to ask what kind of hardware the voice commands run on locally, so that the response time is so fast?
How powerful is your computer?
I use Nabu Casa. That makes this so fast.
Hi, thanks for this interesting video, but I'm encountering an extra step. I'm using an atom echo and he is asking for an encryption key at approx 5:42.
He shows the following, "The ESPHome device Atom Echo enabled transport encryption or changed the encryption key. Please enter the updated key.
You can find it in the ESPHome Dashboard or in your device configuration." Can't connect to ESP. Please make sure your YAML file contains an 'api:' line.
Could you please point me to a solution
just ignore the remark, it is working now without the encryption key
I've got a very specific question. If you setup an ESP32-S3-BOX in a specific area (for example my "Office"), and you also assign it to this area in Home Assistant. If you then use the command "Turn on the lights", does it only turn on the lights in the assigned area (in our example the "Office" lights)?
Yes, I believe it works like that.
use Home assistant and Chat gpt assistant 👍
Does anyone know how this handles voice from across a room with low level noise ie real life situations ?
Aliexpress was asking 56 EUR plus shipping. Bezos arguments aside, I can by two Echos for that money.
True story...
Do you know if there's anything that needs to be done in order to dim the lights? My lights turn on and off fine but when i request to dim to a percentage it doesn't understand. I'm using an Atom echo. Thanks
Mine works for example with Set to 20 percent
I got the atom echo with a view to getting rid of google, but the quality of these things are not even in the same ballpark as google, it will take years of development to even consider comparing them.
Hey, thanks for sharing your experience! It's true that the Atom Echo has some ground to cover before it can compete with Google. But hey, every journey starts with a single step, right? Let's hope for some exciting updates in the future!
Yes, it's a fun/torturous journey haha. from installing ha to atom echo to frigate, nothing sails plainly, but what a sense of achievement. I have my own netflix now, if only I could have my own spotify. Is there anything that can, say....'hey junkie play music', I can't find anything that will just play my server music via voice control :(.....thanks for replying, you make great videos...@@SmartHomeJunkie
Is there an alternative to this ? Faster / bigger screen ?
Not that I know of...
I run HA on a laptop, is there a way to use the build-in microphone for this?
Hmmmm... Not that I know of currently.
Look into "Assist Microphone". It will work with an USB microphone, I am not sure about built-in.
sir home assistant code link plz share
Es caro y consume muchos recursos de la Rpi. Tienen que mejorarla muchisiisimo y bajar el precio
Can I use my phone also for calling the HA?
I think no...
On Android you can change the default assistant to HA Assist. But only after interacting via button. Not a wake word.
@ I think Wake Word integration will be added over time. I've read it somewhere in a forum, but I don't know where anymore.
@ Indeed! Also works on a Macbook like that.
Fisher Stevens doppelgänger
That is unbelievable, 55€ for an esp32 with a small display and plastic case!
You get a raspberry pi for that amount of money...
Awesome. Just wish HA was easier to use on Windows. I don't want to use a Docker container or VM, or a Linux-on-Windows system, and I don't want to fight with compiling the shit myself. Native support please!
Are you asking that they migrate the whole platform off Linux?
@@TheBroz migrate? No. But the platform isn't doing anything unique to Linux. You realize projects can be coded for multiple platforms in a single code base, don't you? Since there's no/very little graphics related things, I don't imagine there would even be all that many OS checks needed.
It's out of stock 😢
Oh no, I'm sorry to hear that it's out of stock! Keep an eye out for restocks, they might be back soon.
Good Video 👍 leider ausverkauft 😡
The tutorial also works for the Athom Echo. Maybe you already own one of them?
Yes, tow of them
@@manfredzimmer1199 Ah great, than you can use the tutorial for those two. I have it running on an Athom Echo as well.
The local option is too slow. This is not pratical at all at this moment.
How slow is it? I think my Google Home is slow as well... Using Nabu Casa it's lightning fast.
@@SmartHomeJunkie slow like 45 seconds to reply. In not on a Raspberry, I'm on a decent mini PC. With Nabu is fast, yes
Try vosk addon from Mike . More fast and accurate than faster whisper.
Can't find that@@Krull56
@@Krull56 Thanks for this suggestion!
I can't understand him.
LOL, I feel you! I have to replay my videos like 10 times to understand what I'm saying. But it's worth it because my content is amazing! :D
@SmartHomeJunkie do you know if it's possible to assign which audio device for the assist pipeline tts responses the audio plays through? I am running a mimic3 server on a different machine with marytts configured to talk to it and I have several audio speakers including a hass agent and sonos speakers I would love to use to play the audio through (instead of the device listening for the voice commands)