The Assist instructions just say go to assist and start talking. But mine only allows me to type in what I want to say, I don’t get a voice option.

What am I doing wrong?

  • @[email protected]
    link
    fedilink
    English
    9
    edit-2
    7 months ago

    You need to install the Assist Pipeline.

    It is fairly straight forward if you can use the add-on manager, but is a little different in docker and there is very little official documentation. If you are using docker just let me know and I’ll share my setup.

    • @[email protected]
      link
      fedilink
      English
      6
      edit-2
      7 months ago

      For those asking, here are my Docker Compose files for Piper (text-to-speech), Whisper (speech-to-text), and OpenWakeWord (allows the use of a wake word with things like the Atom Echo)

      Piper

      version: '3.3'
      services:
        wyoming-piper:
          container_name: wyoming-piper
          image: 'rhasspy/wyoming-piper'
          ports:
            - "10200:10200"
          volumes:
            - '/docker/containers/wyoming-piper/data:/data'
          command: [ "--voice", "en_US-danny-low" ]
          restart: always
      

      Note: The voice can be changed depending on your preference. Listen to the possible options here.

      Whisper

      version: '3.3'
      services:
        wyoming-whisper:
          container_name: wyoming-whisper
          image: 'rhasspy/wyoming-whisper'
          ports:
            - "10300:10300"
          volumes:
            - '/docker/containers/wyoming-whisper/data:/data'
          command: [ "--model", "base", "--language", "en" ]
          restart: always
      

      Note: The model can be changed depending on the power of your system. See possible options here.

      OpenWakeWord

      version: '3.3'
      services:
        wyoming-openwakeword:
          container_name: wyoming-openwakeword
          image: 'rhasspy/wyoming-openwakeword'
          ports:
            - "10400:10400"
          volumes:
            - '/docker/containers/wyoming-openwakeword/data:/data'
            - '/docker/containers/wyoming-openwakeword/custom:/custom'
          command: --preload-model 'ok_nabu' --custom-model-dir /custom
          restart: always
      

      Note: Even though the model for the wake command is set in the Compose file, this can easily be changed in your HA Assist settings.

      After they are up and running, you can just manually add the Integrations to HA which will create the Wyoming Protocol Integration. Then go back to your Assist settings and they should be available in the relevant dropdowns.

      I hope this helps

    • @DaveOPA
      link
      English
      5
      edit-2
      7 months ago

      Thanks! Interesting this isn’t mentioned on the other instructions, but now that I’ve set this up I can talk to it.

      Thanks again!

    • @[email protected]
      link
      fedilink
      English
      37 months ago

      I am about to set up HA in docker, could you let me know what problems you came across and (presumably) ironed out?

    • @[email protected]
      link
      fedilink
      English
      27 months ago

      It might be a good idea for a post. I run HA in docker and haven’t dug into the documentation yet.

    • @[email protected]
      link
      fedilink
      English
      27 months ago

      I don’t know that I need a full on guide, but do you mind sharing what containers you used for Whisper/faster-whisper and Piper? There are a lot of options on docker hub and knowing which ones to use will save me a lot of trial and error.