#Voice PE without TTS

1 messages · Page 1 of 1 (latest)

rancid skiff
#

Is it possible to use Voice PE without TTS?
My PE is only doing simple stuff like turning on lights, hence I don‘t need it to answer with „ok“ or „done“.

Whenever I just disable TTS in my pipeline, Voice PE is not working any. It reacts to the wake word but nothing happens. Even the LED circle is not turning on.

This is what I get in the logs:
There were no events in this run.

pearl ivy
dusk dock
rancid skiff
#

So there is no way to just in general disable TTS? This may be a feature request.

pearl ivy
dusk dock
#

Pipeline right now consists of wake word, TTS, conversation agent and TTS. It's four pillars - IDK if it's hard to just pull one out. 🙂

shut wagon
#

You can create a wyoming dummy server that only receives data from HA but does nothing. For example, return an empty file for all requests.

dusk dock
#

Also if you're experienced with creating your own Wyoming implementation, please share as much as you can. I'm digging into that protocol now.

shut wagon
# dusk dock Also if you're experienced with creating your own Wyoming implementation, please...

Based on the already known wyoming-say integration, I managed to create a server for a local (Russian) TTS engine. To be more precise, the work was done by Gemini. Working with chunk processing and other aspects is quite complex for me. If you're interested in taking a look, the result is here. https://github.com/mitrokun/vosk_tts_hass/tree/main/wyoming_vosk. However, you'll need to translate all the comments and descriptions.

dusk dock