"WHO" says Done when triggering an automation with Voice Assist

Hello Everybody.

I had for some time my Voice Assist triggers in the configuration.yaml.
Some things are not so convenient for me that way so I switched in some automations to a trigger “When a sentence was said” (I have it in german, maybe it is slightly different in your english inst)

My configuration is

  • Assist Microphone AddOn with an Anker PowerConf S330 Analog Stereo as suggested earlierdays in some Wiki.
  • Speech-to-Text is the Nabucasa.
  • A Tasmota ESP32 door-motor on my ground entrance which I can open with the sentence “Open the door”.

Everything is / was working fine.

In my automation I have a Media-out stating “Of course, I will open it right now”.
But in the mid of this sentence somebody shouts D O N E !!

Deactivating the automation / trigger in the automation, the result is “Sorry, I couldn’t understand that” which sounds feasible. Nothing else, no DONE!. Nothing.

The thing is: In assist microphone, there is a field “Done Sound*” with the explanation "Path to WAV file to play when voice command is finished (empty to disable). And this is as empty as a field can be.

I restarted, rebuild the add.on, uninstalled-installed, rebooted the server… it still says Done to me. Not only that, it SHOUTS out LOUD on top of any other media-out.

My shared-folder /share/openwakeword does not contain any done.wav.

Any help appreciated. Jörg Hopmann

I have the same problem.

alias: Wetter Heimat
trigger:
  - platform: conversation
    command:
      - Wie ist das Wetter in Heimat
condition: []
action:
  - set_conversation_response: Das Wetter in Heimat ist warm
mode: single

In this case everything works fine. But when I edit the response in this or any other automation and add a sensor, the resposne is replaced with a “Done”. (The sensors are all valid as I have chosen them from the selection of available entities.)

alias: Wetter Heimat
trigger:
  - platform: conversation
    command:
      - Wie ist das Wetter in Heimat
condition: []
action:
  - set_conversation_response: Das Wetter in Heimat ist {{sensor.heimat_wettervorhersage}}
mode: single

In the moment I replace the variable again the a constant word the response is spoken right again.