Script/plugin/docker container to manage custom scheduled reminders for voice assistant preview?

[obtuse summit]

Is there a script/docker/plugin for HA voice preview edition that i can set reminders/intervals where it will speak them out?

[short helm]

well you can do tts.annouce from an automation and target the Voice PE

can use that in an automation or script

[obtuse summit]

Thanks :) Setup a few now. Is there a way to make the LLM speak it out? As that way, it's response will be with a few more words, and possibly different tone.

[short helm]

You would need to make that part of your automation, that's a bit more advanced. You'd need to use conversation.process i think, then use that response as the text for the tts

[obtuse summit]

Haye, tried to make gpt make a script for it, but HA didn't like any of it lol.

alias: LLM Assistant with Dynamic Instructions
mode: single
trigger:
  - platform: event
    event_type: llm_request
    event_data:
      command: "assist"

action:
  - variables:
      assistant_instructions: "{{ state_attr('assist_pipeline.default', 'instructions') }}"  

  - service: ollama.chat  # Calls Ollama integration
    data:
      model: "llama3.2"  # Uses Llama 3.2
      messages:
        - role: "system"
          content: "{{ assistant_instructions }}"  # Uses HA's assistant instructions
        - role: "user"
          content: "{{ trigger.event.data.prompt }}"
    response_variable: llm_response

  - service: tts.speak
    data:
      entity_id: media_player.home_assistant_voice_097f6f  # Your speaker
      message: "{{ llm_response.response }}"  # Uses Ollama's response

  - condition: template
    value_template: "{{ llm_response.response.endswith('?') }}"  # If response ends in a question

  - service: homeassistant.service
    data:
      entity_id: media_player.home_assistant_voice_097f6f
      command: "Change Home Assistant Voice 097f6f Finished speaking detection to next option"

[short helm]

Why not just use the automation builder in ha? 🙂