When my conversation agent is run (Extended OpenAI Conversation) it starts a script which process the user request. The debug from the Voice assistant shows OKEY as the response. Is there any way of changing this - ideally I want to select from a different set of responses like “sorted” / “no problem” / “done” ?
“Breaker breaker”.
“10-4 buddy” CB Radio style?
Affirmative. Should be possible.
Yes, Sort of - OK gets boring after a time so adding variety makes it more interesting. I’ve tried altering the prompt in Extended OpenAI Conversation to say something else, but that hasn’t worked. The “OK” occurs once my script has completed, so I think it has something to do with voice assistant processing. The debug seems to show it.
any thoughts on how?
Your log is missing the intent-progress, which means the script isn’t running natively, like a tool. However, this isn’t a simple script call (processed_locally: false).
Please provide the script code to better understand the situation. And what version of HA are you using?
I’ve only just seen your response so apologies for not following up earlier. This interaction is through a conversation agent (local LM) that has a prompt to pass the user’s request (simplified by the conversation agent extended_openai_conversation_2) through to a Router routine that then deterministically processes the request. We specifically switched off the intent because the Router is processing the intent directly (with further help if necessary from another local LLM conversation agent).
So we get “OK” from the first conversation agent and when the request is processed we get a second OK.
Its not a big deal so if there is nothing obvious, don’t worry, but I would welcome a way of switching off the first OK that somehow comes from the initial extended_openai_conversation_2.
Thanks in advance,