Project Description

This year was declared to be a "Year of the Voice" in the HomeAssistant world. And indeed upstream community added a ton of functionality and integrations to make voice operations possible.

In my hackweek project I would like to learn about this effort, how to enable it and configure my own instance for voice operations.

Goal for this Hackweek

First goal is to learn about voice usage in the HomeAssistant. Learn about Wyoming protocol, Whisper model for speech-to-text translations various other models for text-to-speech.

Next I will need to adapt my instance, which is k8s based HA, to deploy new containers with all the models for speech translations.

Lastly, since I don't have any HA enabled sound devices, I will need to hack it with PipeWire/PulseAudio server to get full working solution.

Resources

• HomeAssistant
• Assist documentation
• Wyoming protocol
• PA integration