Home Assistant Voice - PE - I don't get it

Configuration Voice Assistant
1

Piper, Whisper, Wyoming, addon, integration, I’m lost…

I just got my Home Assistant PE and connected via the android app to my wifi. It is in my HA (HAOS) and as such an esphome device.

So many questions as I cant get it working or I don’t get it. Let me first “show” some observations:

  1. ESPhome has a greyed out “button”

    image
    image1369×1903 271 KB

  2. Wake word works but nothing works after that.

  3. When I restart the setup (link from esphome in point 1. I get:

    image
    image746×989 63.1 KB

but only mls 5809 (low) gives “audio” on the PE… I can’t understand what it is… all others doe nothing. When I select anything other than mls 5809 (low) the blue ring on the PE spins forever.

I guess I have the following questions:

  1. How and where can I see what PE heard I was saying? (Like in Alexa I can see).

  2. Where is the right place to select a proper voice as now I can’t hear anything.

  3. What is the correct way and order to get things going with Piper, Whisper, Wyoming, addon, integration…

Hapy to share and answer any question or exaplanation.

2

Help?

I can’t get anything going…

I have alamp with an alias.

Whatever I say it tells me "sorry I can’t find a device with the name - - - - .

I ask: switch on office light (office light is alias).

I get:
cant find:
a fine fight
offensive light
off an eight
and all sorts of other things…

I speak so very clear…

(examples thanslated from Dutch)

And any response starts with a very unauaible preamble …

hmz… this might be the issue???

image
image1218×828 52.9 KB

3

First, show the Assist configuration settings
Settings - voice assistants - “home assistant”
Also read this

4

image
image989×424 16.7 KB

image
image708×1305 54.3 KB

What more?

5

The information on the website does not explain the nature of the problem. Dutch has one of the best support in the project.
Only the standard tts and stt (which are not directly related to HA) are a problem.
You should try other intsruments for recognition. Local integrations (vosk, rhasspy-speech) or free cloud integrations (azura, groq…) can improve your experience.

6

Really? Would you be so kind to share what settings/integrations/TTS/STT you use?