Thank you so much for this integration…
I work a lot to make communication between HA and N8N, but with this integration it’s perfect.
Thanks a lot again
Denis
Thanks for this integration! I’ve been experimenting with n8n and your webhook conversation integration, and it’s working well.
Currently, I’m trying to run some tasks in parallel using n8n. I’m not an n8n expert and haven’t found a good solution so far.
My goal is to create a webhook entry for the STT process, then extract the audio — so far, so good.
After that, the audio should be converted to text (i am doing this with the google api), and in parallel, I want to use an LLM to determine who is talking to the assistant.
In the first step, it would be enough to identify whether a male or female voice is speaking (me or my wife).
Then, I want to store this information in memory, so that when the conversation webhook is triggered, the assistant knows who is speaking.
Has anyone already done something similar? Or does someone have a hint on how to run multiple tasks in parallel and wait for the latest one before sending a response? 
I am currently working on an addon/integration that does exactly this: GitHub - EuleMitKeule/speaker-recognition: Speaker recognition service for Home Assistant using voice embeddings. Train on voice samples and identify speakers through STT and conversation agents.
It’s not quite finished yet, but not far from.
Not really sure where I left off development last time I worked on it, you could try it out, it might already work.
1 Like
I got it working with n8n 
But it’s very slow; I think this is because the webhook doesn’t stream the audio.
I can’t find any way to stream audio in a request to n8n — it seems that n8n only provides response streaming.
Is it possible, with your webhook conversation integration, to call a different webhook than n8n and stream the audio to that endpoint? So I could write a small python or typescript script and call it with your integration maybe.