Speaker Script

This is an example of a script which can be used to deliver custom responses.

• It uses the speaker closest to the user
• It switches from cloud-based voice to local voice if the internet is down

tts_response:
  sequence:
  - if:
    - condition: state
      entity_id: input_boolean.use_cloud_service
      state: 'off'
    then:
    - action: tts.speak
      metadata: {}
      data:
        cache: true
        media_player_entity_id: '{{ states(''sensor.speaker'') }}'
        message: '{{ tts_sentence }}'
      target:
        entity_id: tts.piper
    else:
    - action: tts.speak
      metadata: {}
      data:
        cache: true
        media_player_entity_id: '{{ states(''sensor.speaker'') }}'
        message: '{{ tts_sentence }}'
        options:
          voice: XYZ123
      target:
        entity_id: tts.elevenlabs
      enabled: true
  alias: TTS response
  fields:
    tts_sentence:
      selector:
        text: {}
      name: TTS_sentence

Notes

input_boolean.use_cloud_service is an input_boolean flag - an automation turns it on when the internet is available, off when it isn't. It is also turned off if Elevenlabs credit has run out.

sensor.speaker is a template sensor holding the entity ID of the speaker nearest the user. There are several ways to figure out where the user is - motion sensors are quite effective, or you can use a tracking integration like Bermuda.

tts_sentence is the sentence generated by an intent

voice: XYZ is the code for the voice being used (so that Elevenlabs can charge you for it).