#voice-assistants-archived
1 messages · Page 34 of 1
Or even just add a door sensor that is supported
Then you may not be able to close it, but you'll know it's open
I'll think about it
Ah yes I have a door sensor on mine works perfectly
Thank you very much! It works now perfectly.
from what I see the tuya "garage door opener" is just a relay and an input for a sensor
trivial to replace
with homeassistant you could also just close the garage door automatically instead of just alerting you.
or both
I got an answer back from my email to nabu casa. Seemed pretty good so I'm posting it here where frankly I see more requests for help than responses. Hope this helps somebody:
Q:
I'm new. I believe I have everything installed and configured on the Home Assistant Cloud side of things but when I'm in my Alexa App trying to create a routine, the "Smart Home" menu item always reports "No Devices Found" in the "When this happens" phase of the routine. In the Add Action phase of building the routine "Smart Home" does offer all my devices and all my scenes and scripts.
When I include a HA script/scene in the action part of a Alexa Routine, it works.
I'd just like to see my buttons and switches in the "when this happens" part of the Alexa Routines. I have several Zigbee and ZWave buttons and door sensors working reliably,
But I also have a light and a couple of plugs that are only controllable via Alexa at this point that I would like HA to be able to control.
Now I have a software-only MQTT switch which I can turn on and off via Alexa routines and the current state of it is tracked by the Alexa App on my tablet even when I flip the switch from the HA web app. So, if I could just include that in the "When this happens" part of an Alexa routine, I'd have everything I need.
So, basically I'm asking if you know what mistake people commonly make to end up in my current situation, and what to do about it?
A:
Alexa only supports some devices for their routines, but there also appears to be an issue when adding them.
These are generally going to be binary_sensors in Home Assistant and would need to be exposed to Alexa, so things like door or window open/close sensors or motion sensors. After exposing them, you'll then see the sensors available when starting from the "when this happens" section, but they will say they cannot be used. For some reason, you need to wait (24 hrs in my case), and then they can be used as triggers. This is not anything to do with your end or ours and it happens even when I add things which are outside of Home Assistant. I don't know the reason for it.
So lights or switches cannot be used as triggers as Alexa only offers those in the action part of the automation. You can however make automations within HA using almost anything as a trigger, of course.
anyone know what command to use to arm the HA alarm , i set it up in YAML with PIN , and is showing up in the home app , but dont really know how to activate by voice ?
Hey all, right now I’m using rhasspy for my voice assistant (satellite setup). I’m trying to move to Stanford’s Almond and Genie. Only issue is, I can’t get genie to link to home assistant. I’ve tried disabling security, I’ve tried linking from genie to home assistant, I’ve tried linking up home assistant to genie using the Almond integration, but nothing works. I’d really, really appreciate any help at all because DuckDuckGo doesn’t yield any results no matter my search terms.
My environment is HA in a docker container, and genie in a docker container. Rhasspy doesn’t have an issue connecting to HA, which is also in a container for both the satellite and the backend.
Any tips on what I should do, or which system I should move to, would be appreciated.
Thank you to whoever reads this.
EDIT: I just realized that I’ve been using the latest tag instead of stable. I’ll give it a shot and see what happens.
EDIT 2: there is no stable docker tag.
OK - figured it out. I didn’t put the almond server on the same virtual network on Docker. Oops. One day I’ll learn…
Moral of the story: if stuff can’t communicate, stuff doesn’t work.
Thanks for the update.
OK, here's the noobish error I was making: There's an ALexa card under the Home Assistant Cloud config page on Home Assistant.
It has a "SYNC ENTITIES WITH AMAZON" button. You apparently need to push that button for the Alexa App to see your newly created scenes or scripts.
I'd been having problems with my Google Assistant integration so I removed it from the Google Home app and when I try to re-add it the redirect window appears but then disappears, followed by the error "Could not reach <name>. Please try again.".
Any idea on where I should go to try and debug this?
What system are you using? Rhasspy? Alexa? Google assistant? Siri / HomeKit?
If you’re asking for the service, at least for alarmo it’s {entity_name}.alarm_arm_away
For disarming you’ll need a data: block with your code in it like this:
data:
code: ‘{code here}’
@errant kite im using google assistant, i can disarm by voice , but activating it doesn't work
Any errors in the log?
hi so i wanna know if someone tried to do something similar
i want my alexa to do a follow up response
exmple :
turn on the tv
tv turns on
whould you like to turn off the lights ?
cuz im tierd to use 2 commands lol
I've been having issues setting up the Google Assistant integration. Despite my home assistant instance being available over my domain name with TLS and following the tutorial, when I attempt to add the device [Test] HAS to Google Home, I get a white screen which lasts a second or so instead of any login prompt, then an error saying "Could not reach [test] HAS. Please try again later." Home Assistant is not added to my home screen. I also do not see any request logs under my reverse proxy mentioning anything referring to auth/authorize, auth/token, api/google_assistant requests. I use Home Assistant on NixOS on a home server. The network between the reverse proxy on another box and my home assistant at home is a private Yggdrasil network. I am perfectly able to access my home assistant instance through the website, and /api/google_assistant, /auth/authorize, /auth/token all return 405: Method Not Allowed when accessed via a browser (as expected). I also do not see any logs indicating anything wrong under Home Assistant. The reverse proxy works just fine.
I have tried:
- Recreating the action within Google Actions
- Changing the box I host the reverse proxy on
- Using an A record instead of a CNAME record
- Adding the port (443) explicitly into the action URLs
- Adding and removing trailing slashes and every such combination thereof (to my knowledge)
- Using a swedish VPN to attempt the linking
- Renaming the .json file I use for the service account inclusion
- Wiping the Google Home app storage and cache.
and probably more.
I have also scoured just about anything I could for any issue someone was having like this, usually everybody gets the login screen when they attempt to link, at least.
As far as I can tell, Google or my phone just simply don't want to talk to the Home Assistant server. There's very little debug information if at all any, I can't even see what that white screen is as a request. This persisted as the result even between recreating the action, changing the box for reverse proxying and so on. The phone in question is a fully up to date Pixel 6.
Can you do the integration on a computer with the f12 console open so you get better debug infos
What does doing the integration on a computer entail?
Does that mean using the test section in Google Actions?
if so, yes, one moment.
That honestly doesn't provide any data.
For example, when I click "Talk to HAInt" in the suggestions on the Simulator, the response is "We're sorry, but something went wrong. Please try again."
In the console, all I see is CORS complaints, but not to my website, to play.google.com, about logging. :p
I meant try to do the setup process from scratch on your PC
I mean, when I created the action / integration stuff, I was doing all of that from my PC.
Perhaps also with debug logging on HA
but, mmm, I don't see what recreating the whole action again would do, exactly
I'll drop logging level from INFO to DEBUG
It would give debug info that we could use to see if HA is communicating with Google, and vice versa. If nothing shows up when you attempt a connection then it’s either a mismatched port forward or something on google’s end. If there is communication it should give info on exactly what went wrong.
I mean, as I said, the actual access logs on my reverse proxy show that there is no access from Google.
There is no port forwarding, I use a tunnel (specifically yggdrasil) to a box on Oracle Cloud Infrastructure, which reverse proxies it from that box on 8123 to a domain that I control and provides TLS for it. (specifically, https://home.kittywit.ch, which as you can see, works just fine.)
as far as I can tell, the reverse proxy is configured just fine, as I use the external address exclusively for my Home Assistant Companion stuff, and that's never been a problem.
I think it must be something on Google's end, because I've literally done the whole reverse proxying onto a different cloud provider too, BuyVM
and that also had the exact same results.
White screen for a second then "Could not reach ..."
Hmm, wonder if it’s a firewall on your tunnel or on your box’s side that’s blocking google
So, on the whole side of that thought, I tried two different cloud providers
and the only port involved here is 443, you know?
And if you can access it over HTTPS, so theoretically should anyone else (and I checked for rules that would do Geo-blocking, but I don't do that.)
so there's, realistically, no firewalling that should prevent it on my side
that's what has me so stumped, really.
Try adding a trailing slash at the end of the fulfillment url
When you were setting up your action, did you uncheck the “google to transmit clientid and secret via http basic header” box?
Do you have an ipv6 connection?
Mmm, so...
My home connection is IPv4+IPv6, the tunnel is IPv6 only, the end box is both IPv4 and IPv6.
Hm.
Make sure you have a AAAA record for your server then
It does.
or, it did, when it was a CNAME, sec
I'm not really sure why that would affect things though
It should just fall back to v4 and then the reverse proxy should be fine occurring over v6.
Dunno. I’ve had my fair share of ipv6 headaches lately.
My thought was that if google is attempting to reach it over ipv4 and v6 is the only one that’s available, perhaps it’s choking on the 4to6 conversion
hm
but I mean
either way it'd still poke the server with requests
and I'd have logs?
anyway I'm gonna go try and do Dynamic DNS
because my head is going to explode if I don't just do something to replace this solution and see if that works.
alright that's set up
time to test it
once again, the same result :o
I can only assume that this is a result of the vpn / tunnel. Google is up, your home assistant is up. Search the forums and you’ll find posts where people who use a vpn or tunnel have issues.
Please calm down
There has to be something wrong with the link or with config, otherwise we’d be experiencing a flood of forum posts or messages here of other people having issues
That's true.
I'm contemplating whether it's websocket issues.
Not entirely sure, though.
Especially since there’s a whopping 20.2% use rate among everyone who has analytics enabled
Yeah
So, just in case, I’d really appreciate it if you could remove and add from scratch the google assistant integration, and restart your cloud configuration from square one.
Following the instructions here
https://www.home-assistant.io/integrations/google_assistant/
It’s not guaranteed to work but it’s the “turn it off and back on” approach
Alright
same result .-.
y'know, the tutorial mentioning credentials without telling you to generate a service account token provider and downloading the JSON file and like, saying to hit "Add Scope" twice giving you an empty field is really weird :o
also it seems like a good portion of it is entirely out of date .-.
and the same result after doing it again twice
SO!
I took a screen recording of what happens.
That white screen?
At the last frame before I get a could not reach error, it shows me:
"Link an account to control HASS on any device"
"This link will be available on devices or services where you've signed into your Google Account."
"You can change this link in your Google Account."
"Cancel" "Continue"
but!
that exists for a SINGLE FRAME
does that provide you any information of what's going wrong? :p
ah, see
that isn't an option anymore
it doesn't exist
I literally can't select that o:
I literally can't do it
Well
There’s your problem I think
I’m sorry, don’t know how to help you re-enable it
It doesn't have that selection for me, I would send a picture but that's not allowed :o
Dm
so!
i talked to someone far smarter than me about this
@errant kite to solve this, you need to go to Deploy -> Directory Information, fill the whole lot out, go back to your actions and resave the action
and apparently that makes it correctly call to the oauth endpoint
i'm gonna go test that.
it worked.
somebody go and update the tutorial to include this information, thanks, and like, completely redo the images?
@verbal nova thank you so much (and the person who helped you), I've been stuck with the same loop for a while and going through to fill that out has got mine working!
ah :o
i'm glad it helped you ^^
that tutorial on the website honestly needs a revamp .-.
I hadn't had to do it before, I was trying to re-add to a new HA after my last instance died (SD card gave out) so I think that something about the Google app hadn't refreshed, so it probably had something cached somewhere that I hadn't been able to force a refresh on
I've updated the docs with this additional troubleshooting step - https://github.com/home-assistant/home-assistant.io/pull/22293
i would've probably redone the whole tutorial at this point (since it is completely incorrect for new-comers), but mmm o:
posted a reply >~<
i have, apparently, a lot of criticism for the tutorial
if i didn't need to go to sleep and i didn't have work tomorrow i'd write a replacement
(i'll do it tomorrow after work, likely.)
I didn't follow the tutorial originally, I following a video tutorial, I just went back for when there was troubleshooting needed
is it the video tutorial where the guy exposes the private key and a bunch of other stuff for his service account alongside his home assistant url and the project id?
because I enjoyed that tutorial o:
Hi so amm
I wanna use the same command on a different echo and have it turn on/off different devices
@thick flame I don't think that's possible. Since home assistant has no idea from which speaker the command was activated. Maybe you can try something with rooms in alexa itself.
Well I kinda did it so my echo runs off a different Amazon account
And I can just control the devices using home assistant
I'm not sure what your question is exacly
Hi. Is there any way I can combine a jingle with a TTS announcement and play it over the Google Home speaker?
Imagine being at a train station. You usually hear a short jingle (to get your attention) and then a tts message saying when the trains arrive and etc.
Is a similar thing possible with Home Assistant or NodeRED?
Thank you for the answers!
Hi, could you tag me or dm me the link when you will update the tutorial ?
Thank you a lot
I'm using Rhasspy with home assistant but my problem isn't home assistant related I'm just not sure how to accept an input that isn't pre-defined inside of it and couldn't figure it out through the docs my guess is it has something to do with Language Model Mixing but it seems rather complicated for something I figured would be easy enough
(create | set) [a] reminder [named <name>{reminder}]```
I'm looking for some syntax like this where I get a slot called reminder that tells me what the name of the reminder is where <name> can be anything and not something pre-defined then afterwards I was going to add another option for the date/time
Hi, I'm new to smart homes and would like to get my Alexa fully sorted by the end of this week
Biggest use cases:
- Can Alexa order all the items off my Amazon lists?
- Can Alexa ask me a question at a specific time - and trigger an Integromat automation?
This channel is about the use of voice assistants with Home Assistant, not for general questions about using voice assistants.
I mean, everywhere I ask - I get the same response "you shouldnt ask that here" - and get told to ask in XYZ place - I ask there and get the same response
Where do I get an answer then?
anyone using the Alexa integration and find that announce notifications don't work anymore? I can use tts ones, but announce shows it on the screen, etc... which makes them a bit nicer
I have an echo show 5 and used to work fine... think it quit working with the new HA update
Be careful - they don't allow Alexa discusssions here
they do if it's in relation to HA 😉
don't get all bent out of shape lol
Everyone here says you can do it "with Home Assistant"
I come and ask here they say - you can't discuss it here
I asked: - Can Alexa ask me a question at a specific time - and trigger an Integromat automation?
They answered: "Yes, but through Home Assistant"
I came here to ask about it and was shh'd
check this out: https://youtu.be/cCw0vWuhjnY
it's exactly what you want, and was shared in your reddit thread lol
well except idk anything about integromat
doesn't seem like that has anything to do with HA 🤔
maybe ask one of the people who had responded in Reddit?
Yeah, been exploring his vid + github
Just so overwhelmed atm
Thanks for the response 😄
Has anyone had any luck using squeezelite and google home together? I've looked into lms player from ynot, but it's quite outdated
Are you running a squeezebox server / Logitech media server?
Yes I am
Cant seem to figure out exposed domains with the google assistant integration, its showing all of them (including the "set top boxes" google hubs and minis I already have listed in google home lol)
project_id: home-assistant-xxxxx
service_account: !include /google_assistant/service_account.json
report_state: true
exposed_domains:
- light
At the moment I just want the lights added
Tried saying "sync my devices" as well as unlinking and re-linking
This LMS plugin https://github.com/philippe44/LMS-to-Cast claims to let you use google home devices as squeezebox players. I haven't tried it.
I'm having what seems weird behavior to me. I setup a command_line cover device (https://www.home-assistant.io/integrations/cover.command_line/)
which seems to work fine in HA. However when I look at it in google home (or try to use Google Assistant) it shows as offline. (Technically Google Home constantly tries to connect, / re-connect to the device but always fails. ) All of my other devices work fine in Google home/ Google Assistant. Here is my entry in the yaml:
cover:
- platform: command_line
covers:
master_bedroom_curtain:
command_open: curl http://192.168.12.56/position?move_percent=0
command_close: curl http://192.168.12.56/position?move_percent=100
friendly_name: Master Bedroom Curtain
Anyone have any idea what could be causing this?
I figured it out.
I needed to have a "command_state" entry as well...
Hi all. I've got Alexa Actionable Notifications set up and working correctly, however, I've noticed that they don't work on Amazon Flex devices. I've run the same action (changing the appropriate ID each time) on 4 devices. It works fine on Echo 2nd gen, 3rd gen, and the Echo Input, but it simply does nothing when run on the Flex. I've noticed that the Echo Input identifies as a media_player.deviceName_echo, just like a standard Echo (as opposed to media_player.deviceName_input), but a Flex identifies as media_player.deviceName_flex, so I'm wondering whether that has something to do with it and whether there is a workaround for this? Thanks.
hello - I am trying to use an echo show to display a message based on a trigger. I can get this to all work thanks to the excellent media player integration using notify.alexa_media. The issue is that it returns to the home screen eventually and I was hoping someone has figured out if there is a way to change this timeout or to keep the message until dismissed by the user. Thanks in advance for any suggestion or guidance.
With the new echo 15, you can do something like the following: service: media_player.play_media data: media_content_id: add sticky note someone is at the door media_content_type: custom target: entity_id: media_player.echo_15
Oh, neat!
Does it work?
it works fine
and it stays on the screen until it is deleted or another sticky note bumps it off the screen depending on what other widgets you have
Did you get this fixed and if so how?
Hi. I recently made a simple subflow in node red, that plays a jingle and then a tts message on my Google home. But sometimes tts isn't prefect and sometimes I would prefer to for example make a manual announcement remotely. I mean like literally speaking into the mic on my phone and the sound being played on the Google Home speaker. Is that possible?
Question.. Using Alexa devices with Nabu Casa. Can I send a 'door chime' or sound to Alexa if the door sensor opens?
Seems like the only way I know how to do this would be to create a template light for the door sensor and on open, create a routine within alexa to say something like, "Front door is open".. Thanks
Is there any way to make my 3 speed fans ergonomic to use in Alexa?
Right now, they are exposed only as percentages. So I can turn on, and turn off the fan, but "Set bedroom fan to medium" simply won't work. I instead need to "set bedroom fan to 66%" or similar. Which is obviously dumb and annoying for a three speed fan.
I suppose it might be possible to wrap each fan in a template fan to be exposed instead, and make the template fan not support percentage, but only support presets?
Anyone know a good speaker/microphone to work locally with Almond?
Or honestly just a good microphone for it
Are there any guides for setting up a fully self-hosted instance of Genie? Looking into it it looks like the NLP is the part I want to self-host that's not normally. I've got some spare server hw and a spare GTX 970 for it.
I did too several attempts with Rhasspy and HA side the intent need to be “programmed” when you want it to be working.. anyhow the [conditional list ] can’t be in <name variable> you can make a list or define a slot there [named, called, for] <name> … and later check the Jason coming out.. on the side of Home Assistant you then need to decode what you sent from Rhasspy and it is a little.. difficult but not impossible.. if not wrong I did it once and if I found the code.. will send it to you
It is an interesting item.. I guess Alexa.. don’t really understand “medium” speed.. but 1) you can program alexa to react via them app so stat when the utterances are “medium” you go 50% or “fast” 100% 2) Depending from the integration you use.. you could define “template entities” and expose them.. but.. it will take a while before Alexa get on page 😉
Nah Alexa supports it just fine. The problem is that the internal entity to alexa mapping configuration is insufficient. The devs moving fans to percentage based speeds without adding in in such configuration to enable me to set up named speed presets, is the real cause here.
I have set up template fans (based on https://community.home-assistant.io/t/add-support-for-named-speed-lists-in-the-fan-integration/306970/22 but with the values tweaked to 33 67 and 100) that I now expose to Alexa, and besides her calling the speeds "presets", everything works just fine.
In theory, you need to run the stanfordoval/almond-cloud:latest docker image with the "run-nlp" argument with a correct config.yaml and secret.yaml mapped into /etc/almond-cloud/config.d/.
But you would still be relying on cloud access to download the exact match file from s3, and the actual speach to text and text to speach makes use of microsoft's hosted cognitive services, which of course, they charge for.
Looks interesting, but I'm not going to be using a Pi for it. I ordered a Blue Snowball mic for it I'm going to try.
Welp have fun, don’t know how that mic performs far-field, but it works ok near-field
I’m worried that once you crank up the gain on it the Signal to Noise ratio will really hurt processing
I’ve wanted to try and have a good enough server to run it, but haven’t had the time
Anyone having issues with Alexa <audio>? When I try playing an mp3 I get "Sorry, I'm having trouble accessing your Simon Says NA skill right now".
Here is the service YAML I'm trying. The soundbank line works just fine. I've also tried the ffmpeg line on Amazon's docs and I still get the same thing.
service: notify.alexa_media
data:
message: "<audio src='https://v3.jovo.tech/audio/gbpADt89-star-trek-computer.mp3'/>"
# message: "<audio src='soundbank://soundlibrary/home/amzn_sfx_doorbell_chime_02'/>"
data:
type: "tts"
target:
- media_player.amazon_device_here
I guess it's not possible any more? Why haven't I found anywhere else mentioning it? https://community.home-assistant.io/t/play-a-local-mp3-file-on-alexa-echo-dot/186187
Figured it out from a comment at the end of that post, I had to add some TLS Ciphers to my forward proxy to get it to work...
Hello Everyone,
Just followed YouTube guide of Lewis (from Everything Smart Home) to set up Alexa Integration. It was successful, except for the last step where Alexa tries to discover devices. My Echo Dot isn't able to discover devices already configured in Home Assistant... Being new to the DIY route, would appreciate any assistance / heads-up over the above mentioned issue.
can i use speakers which are connected via the headphone jack of my Rpi as a media player for google tts in HAOS
would gstreamer work for this
@rancid eagle I don't have Alexa so can't really help with your problem, but generally it's a good idea to read through the documentation on the integration. It's possible that some things have changed since the YT video was posted
Yes! Rhasspy running as Hass.io addon on the PI4 getting voice commands on a shared MQTT server from a PI3 (I’ll get Pi Zero 2 W when they become available) in the office. Voice control without any remote server!
https://imgur.com/a/3gyyBK2
Now I need to work on making the beeps less annoying, the TTS quality better, and add a few more commands other than just controlling the lights
Any one have an idea why Alexia Manage Entities show nothing? It just spins the circle when I click on it. It has been linked as a Skill but cant use anything with my Alexa items. Google stuff works fine.
Hey not sure if we are taking haaska questions here, but I recently setup an input_button which haaska sent to alexa as a scene. However alexa is indicating it's unsupported and there's no option to trigger the scene. The use case is actually to have alexa trigger a flow in nodered by activating the scene. Anyone have any thoughts they might share or things I might want to investigate? Thanks!
Re my last comment, it looks like all automations are treated as scenes so I'll probably go the route of building it as an automation and then trigger that way.
Hey folks - what's the deal with sending commands to Google Assistant? I'm using a container installation - so I don't have add-ons. I've got Google Home/Assistant integration working, but obviously that just allows me to control HA devices from google - What's the best way to have HA send commands to Google Assistant? I read a little about "Assistant Relay", but it looks like that project might be dead?
E.g. I would like to have a script in HA that sends the "turn off TV" command to Google Assistant, since my Chromecast in the TV can turn it off via HDMI.
Having some issues trying to get HA and Genie (almond). Hoping someone else has this working with an HA SSL deployment. See post below for details.
https://community.home-assistant.io/t/genie-error-when-trying-to-add-home-assistant-skill/417946
While I haven’t had experience with SSL, I’ve had no end of issues with the URI issue. I finally managed to get it to work by a) disabling authentication on Almond and b) manually setting up Almond via config file.
HA needs to reach out to Almond, not the other way - at least from what I can tell
It looks like almond reaches into ha for the authentication token but I'm not certain on that either
I did finally manage to get it to work about an hour ago. Even though I had removed everything twice before I gave it another shot today and somehow it worked this time.
A dev in the ha GitHub said they don't know why but it happens sometimes
HA reached out to almond, authenticates, and then deposits a token which Almond can use to reach out to HA. Almond then contacts HA, auths with the token, then asks for all devices and processes what to make them appear as. Finally, HA reaches out to Almond to get a connection and updates Almond whenever there’s a state change. At least that’s how it worked when I finally got it to work (in a Docker setup)
Hi, I configured Alexa to control my dummy vacuum but all I can control is "fan speed" not even names but values 0-2 which corresponds to LOW MED HIGH. Can I somehow config hass to pass more info to Alexa like zones room names? Thank you!
I'm curious how you're handling device names, mine are overly complicated to use wiith voice controls and I'm hoping there's some way of creating a simple alias. If there's not I might put a feature request in for it.
Wow, I just found out about Google Assistants Dialogflow. This is freaking awesome. Never hassle with IFTTT again 😄
I actually couldn’t find a way to alias other than home assistant’s side, by changing the device name
Please do create a feature request as I think it would be extremely beneficial to almond as a whole
Out of curiosity, what are you using it for?
Everything that needs variables. E.g. my vacuum robot is called Dobby. So I made a Dialogflow which reacts to "vacuum X, Y and Z" where X,Y and Z are some of my house's rooms. This triggers an action that's used in home assistant to vacuum the rooms that were said in the phrase. From that one I directly made a second one "vacuum everything but X,Y and Z".
It's called via "Tell Dobbie to vacuum the living room and the dining room".
Other ideas (and things I'm using myself) are volume controls for devices that don't give feedback e.g. IR.
Or you can ask for where person's are and let google tell you the location of a device_tracker.
Neat, thanks. I never really looked into it myself, but maybe that'll give me some inspiration
Just one drawback (which one might be able to bypass) is that google says something like "okay I'm talking to your test app" but for me it's better than not having the ability to do things like that xP one person said you can publish your app as alpha and it should be better then, whatever that means. Maybe it'll say "Okay, I'm talking to Dobby" or something like that. Maybe some day there'll be a better solution or you can disable it. But for now I'm happy xP
It wouldn't be too much of a problem for me as I rarely rely on the voice assistant, but the latter sounds kinda cool ngl
hey guys new to the discord. what is everyones opinion on almond?
IMHO it’s good but it still could be improved quite a bit - try rhasspy or something first, it requires more config but in my opinion it seems to be more stable and responsive, plus is more flexible
@errant kite is the voice responses pretty fluid or sound very robotic?
I'm wondering if there is any integration of FreeTTS, which provides a pretty good quality? https://freetts.com/
I highly doubt it
@tawdry sluice what version of HA are you using?
I guess it is 2022.3.3
Are you using container, supervised, core, or OS?
haos on rpi3
If you have access to another computer that you can leave on all the time, try maryTTS
yeah, but maryTTS in german sounds not really "natural"
I’m not German and have no one to help me compare the different tts platforms available, sorry.
can someone help me i cant add multiple devices to media player line?
- platform: tts
name: ttshomemedia
tts_service: tts.google_say
media_player: media_player.googlehome8979, media_player.googlehome2398
this is part of notify
It doesn't look like it supports multiple entities: https://www.home-assistant.io/integrations/notify.tts/
Hello - what else is required to get the Muse esp speaker to work with Home assistant? I used the web tool - Squeezelite-ESP32 installer but not sure what the rest of the steps are. I found the SlimProto Player but nothing gets recognized.
OK - answering my own question - seems I was in "recovery mode". Once I exited from this mode, the player showed up.
Hello, I am currently using a routine to tell me the state of an entity with:
action:
- service: notify.alexa_media_buro_bb
data:
message: >-
{{ states("sensor.dwd_weather_warnings_current_warning_level") }}
data:
type: announce
method: all
how can I get a specific attribute of this entity get called out instead of the entity?
When I use Google assistant to turn on my light it works perfectly. But when I use Siri (HomeKit) to do the same thing all my devices turn on throughout the entire house. How does Siri interpret the command “turn on front porch” into an action that turns on everything? I mean even my fireplace turns on and my garage door opens and all my motors start up running. How do I debug HomeKit?
I should clarify that the same thing happens for an “off” command. Telling Siri to turn off porch lights causes my refrigerator and sump pump to turn off.
Hi i have problem with alexa, i followed Amazon Alexa Smart Home Skill tutorial, but when i use layla.amazon.com (I'm in EU) in Client ID in account link i get an error "invalid client id or redirect uri" from homeassistant. When i change it to pitangui.amazon.com i got an error "Unable to link HomeAssistant with Alexa". Test discovery in lambda function works fine.
thank you
I did exactly the same today, also with EU server and worked for me directly. did you use your external URL?
Does the emulated_hue still work to control lights from Alexa? No matter what I do it will not find it
HA Device Name Alias Feature Request
https://github.com/home-assistant/frontend/discussions/12738
hey peeps. I have created a template for my dumb blinds with up down and stop and google home recognizes blinds up and down but it doesnt recognize stop instead it thinks the "stop" command is a command for a media player. Any way to make GH recognize stop as a blind command?
` cover:
- platform: template
covers:
master_blinds:
device_class: blind
friendly_name: "Master Bedroom Blinds"
# value_template: "{{ states('sensor.master_blinds')|float > 0 }}"
open_cover:
service: script.master_bedroom_blinds_all_up
close_cover:
service: script.master_bedroom_blinds_all_down
stop_cover:
service: script.master_bedroom_blinds_all_stop `
anyone have any ideas on how to specify which voice Alexa uses for TTS as part of the command? I found a way of setting whisper mode for the announcement ("<amazon:effect name='whispered'>Hello</amazon:effect>") but I would like to be able to set the voice. Documentation online that implies Alexa does support it (<voice name="Kendra">Hello</voice>) but it doesnt seem to work though a service call in HA.
Is there any way of detecting which Google device a command was initiated from. I have 3 google nest display devices in 3 different rooms and I want to execute a "cast the front door camera" script ONLY on the device I initiate the command on which could be one of three. Ideas on how to achieve this? The script casts a lovelace view (which has the camera and a bunch of buttons and sensors on the page) and Im using service: media_player.play_media service
Ignore the idea that it is a CAMERA - the basis of my question is actually how to cast ANY lovelace view on the specific device the command was initiated on?
Is this where I need to use dialogflow? I hope its easier than that.......
My TTS (to Google Assistant) has stopped working and I'm trying to figure out why.
Looking in the logs I find an unknown error with the Home Assistant WebSocket API and it looks like it boils down to a NoURLAvailableError .
But I'm not sure sure what URL it's saying there's an error about. I have a base_url set in the config to the external address (with a signed certificate) for my HA instance, and I've verified that a cached mp3 can be accessed by https://<address>/api/tts_proxy/filename.mp3, so I'm a bit at a loss.
What's next on the diagnostics steps?
Same with Alexa
seems to be something related to nabu.casa
I'm not using nuba.casa, just local install of HA with a public route
Are there any German assistants?
The Google Nest ones support Swedish so I’d imagine they do German as well
@manic dawn
Not sure if this is the correct channel to try this question, but here goes. I'm trying to set up TTS to a couple of nest speakers and hubs that are located on an isolated IOT-network. So I need a firewall rule to let Nest pick up the media file I think?
Anyone know the port?
Atm the Nest hub "pings" and seems to try to play back a file, but it can't find it I think
So to answer my own question, the firewall rule needs to open up port 8123 which in my eyes is not ideal as that also opens the REST-API to my IOT network which sort of defeats some purpose of segregating my IOT-devices. I've created a rule to just open up from google-nest devices (by IP) which I think would at least be better than opening the port to my entire IOT-network
Ideally the homeassistant service would open a connection that way it's establishing a connection from the secure side that the Nest device can piggyback of when it's open.
If I setup google home manually, do I have to keep home assistant public or can it be turned local after the setup?
Hey there, there is something highly unclear to me, I hope some of you can help me get it right :).
I've setup my whole house on Alexa (groups, zones, lights, other devices), and now I want to migrate all that on Home Assistant, while keeping the ability to control all elements with voice, through Alexa.
Let's focus only on lights with the Philips Hue hub for now.
- My lights are setup in zones in Philips Hue
- Alexa has an integration to Philips hue (so new devices are - supposedly - automatically controllable to Alexa
- I have a paid plan for Home Assistant Cloud
How do I migrate all that on HomeAssistant?
Let's say I want to replace an Alexa command "Alexa, turns on the table lights", I want Alexa to fetch "what to do" in Home assistant, not what is configured on the Alexa application.
Can I do that?
I am trying to set up google assistant sdk on my home assistant as a plugin but it says acces denied with error code 403
I read something about someone enabling the test user to make it work if someone could just tell me how to do that
please help me guys
I am trying the same with google assistant sdk and I'm stuck when going to https://mydomain.duckdns.org:9324/, I get the error 'ERR_SSL_PROTOCOL_ERROR'. Help is highly appreciated!
Anyone know where I can get a version of app.js prior to 2.1.5? I am on beta and can’t get local fulfillment to work with the current version. Trying to roll back from 2.1.5 to see if it makes a difference.
The temperature reported by Kitchen_Aircon is 75 degrees Celsius. The temperature reported by South_Bedroom_Aircon is 75 degrees Celsius. The temperature reported by Master_Bedroom_Aircon is 76 degrees Celsius. The temperature reported by Thermostat is 77 degrees Celsius.
Anyone else get the wrong units from Genie/Almond?
Any ideas on this admins?
There are few conditions. First your home assistant instance needs to be available from the public side and it requires a valid SSL certificate whether it's Let's Encrypt or others. Then you have to create a "Home Assistant Skill" by following https://www.home-assistant.io/integrations/alexa.smart_home/ unless you are running Home Assistant Cloud which then you have to follow this. https://www.nabucasa.com/config/amazon_alexa/
In Home Assistant you can decide which domains or entities will be seen from Alexa.
In the configuration file, you'll have to have a config similar to this depending on which region you are using your Alexa app:
alexa:
smart_home:
locale: en-CA
endpoint: https://api.amazonalexa.ca/v3/events
filter:
include_entities:
- switch.foyer
exclude_domains:
- scene
- script
How to setup google home without nabu casa
oh id idnt' see this channel
OOPS
I was asking somewehre else.
So I followed the instructions for integration with google assistant, worked pretty well HOWEVER @floral remnant the Test part didn't work at all, the test always failed
I tracked it down and what APPEARS to be happen is when you tell it to 'Test' the action it issues an ON command then reads the status right away.
If the device responds slowly - like in my case it's a shelly device connected to HA via MQTT - then when it asks for status it is still OFF and the test fails.
And I don't see any way to put a delay in there
So it seems to work correctly ... but the test aprt failed. I wrote a doc error ticket on one thing, the location of the Test isn't as described in that document, Google may have changed the console around - I don't know.
Google keeps changing things so it the documentation might be out of date. I set up mine years ago and it's worked since. If you got it to work it will probably continue to work though 🙂
It seems to work
I can't figure out how to phrase questions for it at times
like i have a temp sensor named Death Star and when I ask for the death star temp it doesn't know
if I ask it for the temperature without specifying it returns all the temperature sensors
so it's a bit odd.
Honestly it's not important to me though.
I have a 2018 Honda Gold Wing that has android auto
Ask the room temperature that the sensor is in
oh !
so in android auto you can have shortcuts
the shortcuts basically send a voice command into google assistant via text (you don't have to talk)
so what I did was
I put a button on my motorcycle that sends the voice command (text) "open the garage door"
If you can run something like Tasker on Andoird Auto you can probably do that directly too, but your way sounds good enough
And remember that you can use routines to create custom phrases. It's a bit limiting, but you can sometimes get around some annoyances with the assistant that way
I am using something named Automate
I think it's similar to Tasker
but the way to make that work is not to my liking.
You have to create a "Flow" that's automatically triggered
then the first step in the flow is "wait for google assistant"
there's no way for google assistant to trigger a flow to run, mainly because Automate didn't implement any intents to do that.
So I couldnj't really come up with a good way to do it.
with Tasker you could connect directly to HA instead of going through the assistant which is probably faster
What actions can tasker take? I'm having trouble figuring this out.
can it for example just send an mqtt message or make an http request?
it looks like it could HTTP POST directly, or run javscript that could do whatever I want
oh yeah that works great
maybe the thing to do is a tasker task that just pokes stuff into the HA api
I don't need full functionality, just a few things
or use that tasker plugin
perfect
oh - sonos finally added their "privacy first" voice assistant to their speakers from I guess what they made from buying Snips - sounds like it does all local processing but I haven't looked into that yet - wondering if it is going to be possible to reverse engineer it at all and use it for other commands...
Been digging in deeper and tried some solutions from the HA forums which seems to have let me get further. Now I get this in my logs: https://pastebin.com/DYrD6Sab I have raised a NC support ticket also.
Hey guys I added all of my avr sources and sound modes to my google assistant but I can now not restart home assistant because it says
Invalid domain media_player.custom_avr_name
Here is the paste bin to my config
https://pastebin.com/szXTYyYy
Can someone please help me fix my configuration file so I can restart Home Assistant
Anyone?
how much 1h of support costs?
Anyone here using Assistant Relay? I set up mine and it disconnects every time, and I need to reauthenticate.
Anyone with a FiOS right had any issues trying to integrate Google assistant to HA ?
Ask it to”set my preferred temperature unit”
Anyone with alexa issues using nabu casa?
I asked in #integrations-archived , but this might be a better place, as I'm just trying to get TTS with Homeassistant working to play on PCs that are around the house. Is MPD still the way to go for that or is there a better solution -- https://www.home-assistant.io/integrations/mpd/
I.e. I want to announce that the Garage door was left open, and HA is closing the door.
I've looked into it, but Assistant Relay doesn't have the Dutch language. I'm also wondering how others are implementing some good reliable TTS with resuming of audio/video. When I use the normal TTS it's not resuming the media. I have 2 Sonos speakers with Google Assistant, 2 nest hubs and 1 nest mini. Best should be to have the Assistant take control of the TTS because it's an 'overlay' of the current playing media right?
My solution to this was having an additional Chromecast audio play the music out of a different speaker.
It separates the two functions completely solving all sorts of issues
Hmm yeah, not really practical but yea
the nest mini is indeed working fine with only the TTS
I've got a relatively straight forward webhook working with siri, you just tell siri the name of the shortcut and siri sends a json message eg "message" : '1234' which my automation uses to complete a url http://...:8200/1234.mp3 which playmedia then feeds to my google group home audio which then plays the track around the house. Is it possible to do similar with a conversation with Google Assistant in the test environment of the action console i've managed to fill the intent with '1234' so the next stage is to send the json with "name ": '1234' to HA and fire the automation. Is this even possible to do with GA?
Genie tells me since 2 days that the certificate has been expired. Does some one know how I can renew it?
Error processing queue item UserInput {
command: { type: 'command', utterance: 'What time is it?', platformData: {} }
}
Error: certificate has expired
at TLSSocket.onConnectSecure (_tls_wrap.js:1502:34)
at TLSSocket.emit (events.js:314:20)
at TLSSocket._finishInit (_tls_wrap.js:937:8)
at TLSWrap.ssl.onhandshakedone (_tls_wrap.js:711:12) {
code: 'CERT_HAS_EXPIRED'
}
Genie says: Sorry, I had an error processing your command: certificate has expired.
To avoid the need for everybody else to guess, this is for Almond
Correct, thank you! 🙂
Or for Hey Ada: https://github.com/home-assistant/addons/issues/2510
Hey can someone tell me if this is possible. I run HA on pi4 integrated with my echos around the house. I have certain rooms like my kids bathroom that have 3 light switches. They’re named Kids Bath Toilet, Kids Bath Shower, Kids Bath Sinks. Can I create a voice name and have HA just turn off and on all the “kids bath lights”?
Make a group and expose that?
Yup, group 'em.
I have home assistant setup with the google assistant integration and nothing else on google assistant. If I don't control anything using google assistant for a few hours and then use it, it has a very noticeable 5 second or so delay before the device is triggered. After that device is triggered, all device controls immediately after work really quick. Is there a solution for this?
I expose home assistant using cloudflare tunnels but I also had this issue when I port forwarded
All the devices on google home have tasmota installled
I see that there have been some significant changes to local fulfilment since I installed it. I will see if these help
wow that made a huge difference, will see if it is fixed
all, I am looking at a way to use Alexa to set a toggle to on or off in HA. I have a routine that needs a voice command to set many toggles to set for like a weekend shift and such
Any idea why my Alexa devices aren't showing up in the media section of home assistant?
My unifi cameras that have speakers are there but that's it
Do you have the Alexa Media Player custom component installed?
There's no built-in support for Alexa-enabled devices
I do not, just the Alexa Media Player integration from HACS
Oh yeah i have that thing
And my Echo devices show up in home assistant
They're just always in standby mode it says
I can also control the volume from Home Assistant and i see the LEDs changing on the echo
So i definitely am talking to it
Which voice assistants can I use the Nabu Casa TTS with?
when you pay the license fee you can use both google and alexa. However you might need something like IFTTT or Voice monkey to push the payload. Depending on your need.
I have Google Nest Mini and I cant play music from it, when I ask it to play music it says it cant do that from here but it can do it on another device ????
I can cast music to it though from my phone
and I can control volume and stop/play from home assistant
Just now.... Me in my kitchen to my Nest Hub: "Hey google, set a timer for 10 minutes"
Nest hub lags like crazy, stuttering
Mini from literally farthest away corner of my house, barely audible to me: "Okay, timer set".
Me, again to Nest hub: "Ok google, show me what timers I have".
Nest hub: "No timers are currently set!"
🤦♂️
I will have to go to the other end of the house to stop a timer in a few minutes, right? Aaaghhh
Yup... It's ringing now. 😐
Sounds like you might need to reset that hub or something
I bought it 2 weeks ago
Might still help 🤷♂️
I have the google voice assistant integrated, the manual way, and wonder if it is possible to enable logging of voice command actions to the homeassistant.log
Try enabling debug logging for that integration to see if it logs anything: https://www.home-assistant.io/integrations/logger/
What google assistant devices have a line output for connecting to external audio equipment? (What is the cheapest device I can talk to and get it to play Youtube Music (I am a subscriber))
I'm having trouble from materials I've read in understanding whether Google Local Fulfillment works out-of-the-box with HA Cloud/NabuCasa or not? I've got a subscription and I have HA running on its own IP, and it shows up in mDNS with the IP as internal_url but nabu casa URL as external AND base. Do I have to do something else to set up local fulfillment? I don't really wanna go the manual Google Assistant route cause it's still YAML-conf and that's super tedious to handle
umm you could just get a nest and a bluetooth receiver
from what I saw nest can cast to bluetooth devices
but dont take my word
check it
okay apparently it can
I've tried that with a Google home mini and found that it was not at all reliable. The Bluetooth connection would stop on it's own all the time
hello
i want to use a local voice assistant and i dont know which one it is good genie or rhasspy if there is some use them give me a feedback thank you
Hi all, I've been trying to follow the steps on the documentation to get Alexa working with the local method for a few weeks now and I can't seem to complete the Alexa skill account linking step. Could anyone point me towards an up to date guide? I have followed the instructions on in the ha docs, but am not able to authenticate the dev skill with my instance.
- When entering ha account details and MFA code into the dev skill account linking in the Alexa iOS or web app I receive the following error 'We were unable to link HomeAssistant at this time'.
- Is there any way to debug port forwarding setup and possibly SSL certificate? For example to test which of the ports may not be setup correctly and to confirm that the certificate I'm using is valid and will be accepted for the Alexa setup?
- I'm using DuckDNS and NGINX.
- Do I need to disable the Alexa skill with Nabu casa before setting up the local one?
Did you get any clarity or guidence?
Hello Guys, which was the command to send a command to amazon echo instead of TTS?
how can i get an alexa device to set a toggle in HA?
I just expose the input_boolean
Perhaps I'm not clear on exactly what you're trying to do
@fallow inlet thanks. ok, but how do i do the exposure? I want to give a voice command to alexa, and have it set a toggle in HA. like if I am in a meeting it will set the toggle in HA on my meeting scene that sets the lighting and a few other items up so i am not interrupted. Then toggle it back when i am done. since HA does not have direct voice access in my environment, and i have a bunch of alexa devices.. just trying to bring them all under one roof.
I don't see the toggle available in alexa to use as a device/action or anything.
I use the YAML config for Alexa via Nabu Casa and I just expose the input_boolean as a switch
There's nothing special
ok, but what type of interface do i need to make? I have the nabu account and i just don't know how to expose the boolean. im kind of new-ish to this
like a sample code would be helpful if you have it
You can do it through the UI, too
I suggest starting with the instructions at the link
Ok. But where/how. I know it is capable, I need the code/ui and where.
I did follow the link, but there must be a way to do it without building a lamda block.
-> Cloud -> Alexa -> Manage entities
this describes that method, and also the YAML method: https://www.nabucasa.com/config/amazon_alexa/
Ok. Let me try again. But i didn't see it last time. Kinda lost.
Alexa found many devices, just not the toggle. If i add the binary will it knock out the others it already has?
I don't know why it's not seeing it. I have this in my YAML config:
alexa:
filter:
include_entities:
- input_boolean.heat
entity_config:
input_boolean.heat:
name: Heat
description: Heat
display_categories: SWITCH
and I can say "Alexa, turn on the heat" or "Alexa, turn off the heat"
do you guys need to re-auth alexa media player with every reboot? this is new every since the launch of the SAML stuff for me
Hey, I'm trying to setup alexa actionable notifications but I've gotten stuck setting up the alexa skill, I tap enable but nothing happens. Please help
can recommend Mark Watt tech Youtube channel I think it is to get it setup
No unfortunately not. I’ve been able to confirm that I have a working certificate for access to HA remotely using NGINX, a non standard port and DuckDNS. I’ve checked for open ports and they seem to all be open and forwarded to home assistant device. Any thoughts on what else to check for?
sounds like you hit all the bases. i am struggling to get voicemonkey to be usable and even sending local media to the echo devices. so we are all in the same pool, just different lanes. lol
Good luck with it 👍
Hi, I've just setup alexa actionable notifications but whenever I respond to one of her questions she replies "You just triggered AMAZON.FallbackIntent."
is there any free alternatives to triggering HA automations with Alexa? Emulate hue has been playing up for me after running pretty solid for 2 years
Hey brains trust - I have a template switch defined as a blind exposed to Google Assistant. How can I find what commands/payloads GA will respond to? I can currently get "Open blinds" and "Close blinds" working but "Stop blinds" doesn't work. Is there any form of document that defines what is valid for a particular domain?
i installed Almond in HAOS and i get a 502 Bad Gateway when trying to open the ui
since google assistant local fullfilment relies on mdns for locating home assistant on your network, how does it behave with multiple ha instances on the network?
Hi, sorry to disturb,
Could anyone explain what this does specifically ?
"You can now trigger Alexa routines from switches, toggles, and buttons without the need to wrap them into a binary template sensor first. "
Extracted from the latest HA release note
Thx
My nest hub can no longer control my thermostat via voice. When I ask my nest hub display to set cooling it says thermostat is not setup. Despite that I can control the thermostat on same device but using touch screen. I can also control thermostat from google minis using voice just fine. Has anyone seen this problem before?
I'm trying to get my Google Assistant to work with HA and I successfully completed every single step described in the documentaion on that matter - only to find "Invalid client id or redirect uri" in the Google Home app when I try to add the skill. The log in HA says "OS Error while looking up redirect_uri: Network unreachable." and "Login attempt or request with invalid authentication from (my phones ip)"
I'm running on a Raspberry Pi 4, using duckdns and nginx to make access possible from outside my local network (and that works perfectly fine via the browser and the android app). I read about an ipv6 issue and disabled ipv6 on the pi but it didn't help. Anybody got any ideas?
Fixed it. I was missing an / in an url in the google actions console. 👁️ 👄 👁️
Anyone know how (or even if) home assistant can filter which devices are exposed to genie?
So...really no one can answer my simple question about that sentence in the last release note ?
Simple answer. This triggers an Alexa routine with a switch, toggle or button.
Yeah, that is the obvious part...I asked for more details
If your comment on your question didn't sound so pretentious you may have received more help.
My suggestion would be to research it a little more and try some things out rather than just asking for someone to do it for you.
Pretentious ? really ?
Who the fuck you think you are ?
Isn't that Discord for help ?
It is. But when you ask why no one can answer you "simple" question... it's a bit off putting
It may be only texted based, but words can depict a personality.
You should look for help... not demand it.
That is because I've asked that question 3 days ago and no one even bothered to answer...I did a bit of research, and finally landed here because there is absolutely nothing detailing this sentence
Well, if even the people who develop HA don't have the answer, I think there is a BIG problem no ?
lol... now demanding HA devs answer your question because you asked it.
wow, you must be really important wherever you are.
I'm not demanding HA dev specifically, I'm demanding litterally anyone who knows what it means
And there's you're problem.
Ok, now I've got a problem because I find an obscur sentence in the release note, for which I search explanations. I do not find anything related to it and come here, expecting someone will have an answer.
Fine
Your problem is that you are demanding a response.
No one here has to do anything for you.
Dude, don't be an ass, what is the point of this Discord if I can't even ask a simple question...
Discord helps bring people from all walks of life from around the globe together. The issue isn't the question... it's your reaction to when no one provided an answer to it. Ask yourself how you think those around you would have reacted IRL if you provided that response.
I don't really know what it means, which is why I did not respond. I just looked up the PR: https://github.com/home-assistant/core/pull/67889
Presumably it means that button and input_boolean entities can now be used in the 'when this happens' part of a Routine
does Google assistant need home assistant to be exposed to the network because of webhooks? Or other things as well?
Well it is a good question 🙂 I guess depends on the implementation, using NabuCasa for example the devices are automatically exposed to.. our Amazon friend. Sometimes it feels a little of interpretation is needed.. I mean if you use no NabuCasa then you might expose the “devices” somehow “manually”, and I believe they mean.. it is now more simple to do so… and by the way.. opposite side apparently is possible to trigger the routine of Alexa from HA means basically.. well you double up the job..
It probably means the switches etc have a new option you can use to trigger an Alexa action by name/ID. Is there an issue number next to it? (#xxxx)
Ok, from what I see in the code, you can force Alexa to detect presence using a button press
Cross posting this from the general channel, Anyone have some suggestion on how to use ADA and RTSP streams for a pseudo voice assistant? Basically looking to do semi simple things like when camera detects movement in a room(and lights are off) ask if lights should turn on
I like listening to the news from Google Assistant as part of my morning routine, but there's more info I want that's only available via HA. I can use the TTS feature to announce that, but can I then get it to run the news command on my Google Home?
I don't know that this is something you would need Ada to do. What I would probably do for this would be to bring the RTSP stream into something like MotionEye that has motion detection support, and then see if there's a way to use that motion detection as a trigger for an automation to turn on the lights.
So I already have motion detection(using UI cameras and presence detection), Id just like a way to do something like when someone is detected ask for instance if the lights should be turned on
Are you using Unifi Protect with your own NVR?
There's a Unifi Protect integration but it doesn't work with the cloud protect platform. But if you have your own NVR, you can use the integration. I haven't used it, and I don't know if it'll give you a motion detected sensor you can automate on, but that would be your best bet.
Yeah I have my own NVR and it all works fine. It just doesnt have STT so doing stuff on replies isnt possible which is what Id like.
OH... I think I misunderstood. I thought you just wanted the light in the room to come on with presence, you're wanting to use the audio stream as an audio input for ADA.
Yes exactly, or to some other voice assistant. Honestly it doesnt even really need to be a voice assistant so much as just some STT service so I can say Yes and it do something
So... I'm just kind of spit balling here, as I feel this is out of my depth. Could you use the VLC add on, have the RTSP stream running all the time, and then loop the audio or do a software mix to where the audio that would be outputing from the VLC stream is then dumped back into the host device as an input?
Thats an interesting thought I didnt consider but maybe. I think the key here is finding some way to take a RTSP stream and convert it to text from there doing anything is easy
I looked at using ffmpeg and it take a RTSP stream and just take the audio, the question is then how to convert that stream to text on demand which I dont have a solution for. I thought HA supported it natively as they had a STT integration but apparently not
Just to update people here I was able to basically accomplish this using azure and some custom code. So whenever a prompt is to occur it publishes on a mqtt topic, and uses TTS to ask a question. The mqtt topic publishes triggers my custom code to start listening on the RTSP stream and forward it to azure to convert to text. And then based on a yes/no the script publishes on another MQTT topic that home assistant is waiting for a reply on
Hello, I'm not sure if I'm in the right place here in this channel, but I have a problem with the Amazon Alexa Smart Home Skill. If I want to authorize the Alexa skill with my Home Assistant instance, I always get the following error message in Home Assistant: "Login attempt or request with invalid authentication from...." and the amazon alexa app tells me the connection is not possible atm. But the functions tests with execution result: successful. What am I doing wrong?
First, I would like to apologize for how many questions I have been asking, I have been trying to do better, but am not quite to the point of being self-sufficient. That brings me to my current question. After having followed the guide to set up a link between HA and Alexa, I found myself unable to access the web interface through either the internal or external URLs, with the internal URL simply saying that it cannot connect, and the external URL giving a 404 through HTTP, or redirecting me to the auth token page through HTTPS. This occurred specifically after I restarted the instance to bring in config regarding devices that should be exposed to Alexa. I did run a check on it before the restart and didn't get any errors, which confused me a fair bit. I have also confirmed that the system is live and have full access to the CLI, as well as that the proper ports are open through an nmap scan. Finally, the only change since the last functional access was the Alexa configuration, which does not seem like something to bring down the entire instance.
And we are up again. It was that configuration causing trouble. Still unable to connect through the internal URL, but can do it through the external URL.
Below is the YAML file. The commented out lines related to Alexa are what I changed to regain access. https://www.toptal.com/developers/hastebin/oruwocapof.yaml
Hi all. Last night I started getting "Null is not responding" when trying to control most of my devices with Alexa. I'm running latest version of HASS and the DIY connection (not using HA subscription service) with everything running in docker. I cannot connect to my Zwave (zwavejs2mqtt) or Zigbee (zigbee2mqtt) devices using Alexa voice commands but they do work fine within the HASS interface. I've also restarted all containers and am not seeing any errors in my logs.
Can someone tell me where I might start looking for the problems?
Hi I just got a zigbee thermostat and set it up with zigbee2mqtt and it shows up in HA just fine, but I'm having trouble getting it to work right with Google Home.
It's showing up in the app, but only shows the indoor temperature and not any setpoints or the status. Has anyone had a similar problem and how did you fix it?
https://cdn.discordapp.com/attachments/708143667453624461/999140686605582416/Screenshot_20220719-222744.png
edit: I think I found the problem, and opened an issue: https://github.com/home-assistant/core/issues/75521
using HA cloud + alexa. I have a room with a mix of Hue bridge connected lights, and Zigbee connected switches (to lamps)
I'm trying to make a singular voice command to turn all of them off and another for on.
I cant seem to group them in a way that alexa can find them to control them in unison. any help?
i tried scenes, but having to say "turn on food room off" sucks for turning off your dining room lights
HA + smart_home custom Alexa skill, can turn on/off devices but can't get the state (Alexa's response "Device does not support this"). interestingly the state is visible in the Alexa app.
while on the other side google can do this flawlessly.
could I use an esp32 with a microphone as a "trigger" for Ada, and let it output from one of my existing smart speakers?
What hardware should I get to recreate a smart speaker? Thinking of an RPi zero w or ESP32 base with a speaker, what should I get for a far-field mic array?
Can I just un-google my GHmini and flash something open onto it and use it for it's hardware without corpo code?