I want to trigger a script via a voice command using Home Assistant Voice PE with a ChatGPT-based pipeline.
Ideally, ChatGPT should respond in a friendly way, confirming that the script is about to be executed, and also include some context-specific information related to the script itself.
Example:
The script starts my coffee machine.
Currently, the script uses TTS to say: "Please place a cup!"
Then ChatGPT says something like: "Alright, I'm making your coffee. Enjoy!"
The problem:
The spoken TTS message comes from the script, while ChatGPT replies afterwards — this feels unnatural and sometimes overlaps.
What I’m looking for is a way to have ChatGPT itself deliver the entire response (including the helpful info), while the script silently handles the background actions (e.g., delays, service calls, etc.).
So ideally, I’d like to pass a prompt or instruction from the script to ChatGPT when the script is triggered. Is that possible?
Thanks for any ideas or workarounds!