I’m currently running home assistant on a HA yellow and have been running voice pipelines for some time. I’m attempting to troubleshoot the lack of streaming the TTS responses. The pipeline works without errors and I do get responses, but when I look at the raw debug they all contain stream_response: false.
I’ve tried this with Gemini for the model and the TTS, HA Cloud for both, as well as my default configuration of the ollama conversation agent and piper for the TTS. It’s my understanding from the docs that the stream response should default to true. Am I missing a configuration setting or is there some advice on how to troubleshoot this?