How to add extra switch to to start conversation with the assist

[witold-gren]

If you need add extra button which you will just click to start conversation with the assist without use wake work just copy yaml configuration below and add it to switch configuration:

switch:
  ...
  - platform: template
    name: Listen Once    
    id: listen_once
    optimistic: true
    restore_mode: ALWAYS_OFF
    on_turn_on:
      - switch.turn_off: use_wake_word
      - delay: 0.5s
      - voice_assistant.start
    on_turn_off:
      - switch.turn_on: use_wake_word  

after that you will see extra button which you can you in your automation.

BTW. Before you start using start_va esphome script you must always disable use_wake_word. This button solve this problem because the do automatically for you 😀

This configuration will not turn off switch when voice_assistant will finish pipeline. If it is important to you to auto disable this switch and enable wake_words then you must add one extra line switch.turn_off: listen_once to voice_assistant configuration:

voice_assistant:
  ...
  on_end:
    - wait_until:
        not:
          media_player.is_playing:
    - script.execute: reset_led
    - switch.turn_off: listen_once    # <----- add this line here
    - if:
        condition:
          and:
            - switch.is_on: use_wake_word
            - binary_sensor.is_off: mute_switch
        then:
          - micro_wake_word.start

[TheStigh]

Hi @witold-gren Question: Do you use the improved audio part (adf_pipline) or the arduino framework?

[witold-gren]

You talk about continuing the conversation, and I would like to completely stop further processing.. I'm not interested in the answer from Intents or any LLM, I just want to know what was said.. 😀 For this reason, I want to stop further processing

[witold-gren]

I know how to make a conversation, but I don't know how to stop processing "requests".

[TheStigh]

Ah, true - wrong code :) I
ll get back to you in a bit.

[witold-gren]

BTW. How you recognise in automation (or custom_conversation) which speaker was use to create question? I have an idea for this by checking recent activity in sensor on my 5 devices... but I don't know if there is a better solution... 😀

[TheStigh]

Using trigger and id's I guess? Then you can check multiple and even use condition, finally use choose in action to fire the correct routine based on the id from trigger.