2023.5: Let's talk!

[DavidFW1960 (David)]

@DevOpsTodd

try to justify this by blaming me for using routines

My routines didn’t break. They all still work after all the entities came back

[DavidFW1960 (David)]

@GSzabados

It is a bit controversial adding the devices this way to any voice assistant, if it has unique_id or not. First all the HA integration should support the unique_id option, by default, being yaml or not.

You can try the expose entity button in bottom right and manually add the entity and then expose it. I picked this up in the beta with my limitless led lights which all disappeared from google but they fixed it and they are all working again now.

[GSzabados (Gábor Szabados)]

@DavidFW1960

they are all working again now

Thanks for that! I will give it a go once updated.

[davidnestico]

Thanks for this, they need to make it an actual card now maybe cuz kind of annoying you have to refresh the page to make it show up, I got kiosk mode enabled so on mobile there really is no way but this to use the assist mic button.

[srstudiostech (S3v3nD34dly51ns)]

Love the idea of this new update, localized voice recognition. Has been a long time coming, however the biggest issue I had with this update is noticing that none of my Alexa integrations were enabled and had to re-enable each device manually. Perhaps an ‘add all option button’ and make re-integration possible. The only thing that really saved my ass in all of this is i had a Hubitat running alongside my HA and all devices were tied to it. Alexa is also tied through Hubitat, so I never lost integration really. I did however plug them all back into HA, kind of a pain the ass really. That’s expected with ground breaking updates though.

[tom_l]

Your other option is to wait or build an ESP based microphone and speaker or VoIP device. Http access is going to be added to the app see:

Year of the Voice - Chapter 2: Let's talk

This is a requirement imposed by browsers. Websites cannot access user’s private information (including the sound in your room) unless the site is served over HTTPS. This is only limited to browsers. Any Assist devices, like ESPHome or VoIP, will not have this limitation. When we add this to the apps natively, it won’t be an issue either.

[rob_woowoo (Rob)]

Likewise, Alexa lost all access to existing devices after the upgrade. I restored from backup to previous version as a short term fix.

[pragdogforshins]

Phew - glad someone else has this, because I have not one single clue as to what to do about it. Downgraded HA back to prior version and all good again.

[sog (Dale)]

My LocalTuya is broken since the 2023.5 update. I see a couple of others having issue with LocalTuya also but no addressing of the issue as far as I could find.

Any news on a fix or workaround for localTuya? Thanks in advance?

[blackie333]

Tried to install Whisper add-on on latest 2023.5.2 but after changing language to sk (with default tiny-int8 model) the add-on started reporting “This add-on has no configuration” and I can’t modify config anymore or start the add-on.
Add-on uninstall and re-install did not helped…

Add-on log seems fine:

s6-rc: info: service s6rc-oneshot-runner: starting
s6-rc: info: service s6rc-oneshot-runner successfully started
s6-rc: info: service fix-attrs: starting
s6-rc: info: service fix-attrs successfully started
s6-rc: info: service legacy-cont-init: starting
s6-rc: info: service legacy-cont-init successfully started
s6-rc: info: service whisper: starting
s6-rc: info: service whisper successfully started
s6-rc: info: service discovery: starting
INFO:__main__:Downloading FasterWhisperModel.TINY_INT8 to /data
INFO:__main__:Ready
[07:17:31] INFO: Successfully send discovery information to Home Assistant.
s6-rc: info: service discovery successfully started
s6-rc: info: service legacy-services: starting
s6-rc: info: service legacy-services successfully started

P.S. After host reboot and removing OpenAI pipeline the add-on has started and config is working.
But now in the HA Pipeline I can only choose English language instead of Slovak which was configured before Whisper add-on installation.
Seems that even Whisper config now uses Slovak Base model the HA Pipeline refuses to select other languages except English (my languge in HA general settings).

Would it be possible to use my ESP32 based “AtomU ESP32 Development Kit with USB-A” device with a microphone but without any speaker for local Voice Assistant? Already have an MPD speaker in the same room for playback of Google TTS messages so I hope it could be combined, but not sure how to assign audio devices to the pipeline stages…

[nickrout]

But if you’d read the thread you wouldn’t have needed to downgrade.

[nickrout]

Isn’t Local Tuya a custom component? If so, this isn’t the place to ask for support.

[Gilean]

Not sure if this thread is right place to ask, but I have been wondering for a long time about this HA Voice thing that promised “you can control HA using your own language!”. So, my native language is Finnish, and I have both Alexa and Google speakers that understand english but not finnish (Google text-to-speech speaks quite fluent finnish though). So, I was wondering if it was possible to use my existing smart speakers (Alexa or Google Nest Hub or both) with HA (which has Finnish language support) so that instead of activating Alexa or Google, I could speak Finnish to either of those speakers and then HA would handle the rest?

[Viking]

After upgrade to 2023.5.2 HA complains about failure to load pyatv:

Logger: homeassistant.util.package
Source: util/package.py:107

Unable to install package pyatv==0.10.0: ERROR: Cannot install pyatv==0.10.0 because these package versions have conflicting dependencies. ERROR: ResolutionImpossible: for help visit https://pip.pypa.io/en/latest/topics/dependency-resolution/#dealing-with-dependency-conflicts

I did not get this error message with 2023.5.0 or 2023.5.1 or any other previous version.
(I’m running HA in a Docker container on a Linux machine)
During firts boot after updating HA, the error message also contained a recommendation to upgrade pip, which I did: pip install --upgrade pip and I’m now at pip version 23.1.2 and Python 3.10.

[Twit (Simon)]

You can either use the original ttyUSB0 but run the risk of the port changing

or use by-path instead of by-id, my Zigbee USB looks like this /dev/serial/by-path/pci-0000:00:14.0-usb-0:4:1.0-port0

I have used the by-path and it works fine.

[TimoJ]

2023.5.2 still shows these unnecessary menubar chevrons that first appeared in 2023.4. Is this now a “feature” that will be there forever?

[mirekmal (Mirek Malinowski)]

Thats the question I asked some time ago for Polish. I’m in exactly the same situation as you. The answer, unfortunatelly was no, it is not possible. The problem is that neither Google or Amazon provides API to either convert speach to text and return it back to caller, or to record just voice and return it as audio file back to HA. Unless something changed in meantime the only solution, I’ve seen some users were considering, is to hack these devices with ESP to use good mics arrays and/or speakers to build essentially new device with old hardware.

[makai]

Is LocalTuya up-to-date?

[makai]

Are you using the custom component or the official integration?

[Gilean]

That’s bummer But I feared it could be so… Thanks for explanation anyway! To me it seems as if HA Voice system thingy was meant for people who carry their phones with them all the time and then speak to HA via mobile phone. But for me the only thing I would need voice is when I am for example laying on the sofa and would like to turn a light on or off or something like that. My phone is somewhere but Alexa / Google speakers would hear me…

Maybe I will look into hacking my Alexa speaker if that was possible, I don’t want to mess with my Google Nest hub because that has display and everything, it’s easier to sell it away…