I’m pretty new to HA so am probably missing something pretty simple. I’ve followed several different videos on how to setup voice assistant. To the best of my knowledge everything is set to default, and the openwakeword entity is showing up in the Wyoming Protocol services section. When I go to settings>voice assistant and click to modify the existing HA voice assistant there is no Wake Word option below the Text-to-speech section.
If I edit the entity it shows as Unknown and says it became unavailable yesterday. If I click the gear icon, it shows as being enabled. Disabling/enabling does nothing to change the status. Neither does restarting openwakeword (or even reinstalling it) though the settings>add-on.
This is a far as my troubleshooting had gotten. I’ve tried a bit of Googling, but must be missing simalar issues.
Always start by reading the docs, this will give you an understanding before attempting to set something.
this will help
Be aware that although videos are very useful they can quickly become out of date. So be aware of this, read the docs before following a video.
It may be even worse with voice assistant stuff as it is and has progressed fast over the last few months.
It sounds like you have done stuff correctly, but there probably is something not quite right, I can’t spot it from what you have written. So A couple of questions.
What hardware are you intending to run this on? What installation do you have, and what hardware are you running it on? This may help other find your issue.
I’m more than aware of the pitfalls of outdated Youtube, that’ one of the reasons I’ve watched multiple videos, and tried to stick with more current versions.
I’m sure the docs would help, but they feel a bit overwhelming to a newbie as there are so many different paths, options, links. I’ll try and wade through them…
Here’s what I’m using:
Core 2024.12.2
Supervisor 2024.11.4
Operating System 14.0
Frontend 20241127.7
I’m running HA in Proxmox 8.3.1 on a GMKTec G5 (8 x AMD Ryzen 3 5425U with Radeon Graphics (1 Socket)
That’s a very weird design decision to hide it. I thought I was going mental because I could configure it for one of the assistants (because it was configured when this was a default setting and always showing) but not for any new one I created.
thanks for the hint with the 3 dots in the corner. Unfortunately, it seems that it is not possible to use the activation words of the installed engines in the Voice Assistant of the nabu case? I have selected porcupine1 in the Voice Assistant. But the other options are not in the ESP. Or am I doing something wrong?
See photos.
I too am having the same issue with my Voice Assistant PE. Spent the week after work tweaking the custom wake word, and now the weekend is around I have run into this road block.
Edit: Also no Edit in Yaml option to force the option either.
I was able to follow the microwakeword notebook to train a custom model, create a json file based on the okay nabu one, and then take over the device with esphome and added my json file to the mww models section