Is there a way via automation to change speech-to-text or text-to-speech or conversational agent of an assist? This way I could use local voice recognition and llm while my computer is on and use cloud systems while it is off
Create two pipeline and switch between them.
alias: voice-agent_change
description: ""
triggers:
- trigger: conversation
command:
- Смени[ть] режим
conditions: []
actions:
- if:
- condition: state
entity_id: select.wsat_assist_pipeline
state: Home Assistant
then:
- action: select.select_option
metadata: {}
data:
option: AI
target:
entity_id: select.wsat_assist_pipeline
- set_conversation_response: Агент изменён на сбер
else:
- action: select.select_option
metadata: {}
data:
option: Home Assistant
target:
entity_id: select.wsat_assist_pipeline
- set_conversation_response: Основной агент выбран
mode: single
1 Like
sorry how to create two pipelines? You mean more than one assist? I don’t have any entity starting with “select.”
Aѕsiѕt is a voice module in HA in a general sense. Inside it you configure pipelines(settings - voice assistant).
The selector is the entity of your satellite
1 Like
I see, thanks