synesthesiam / homeassistant-pipeline

Websocket client for Assist audio pipeline

Assist Pipeline Client

Basic websocket client for an Assist pipeline.

This client:

1. Authenticates with a Home Assistant server
2. Runs an Assist audio pipeline
3. Reads raw audio from standard input, sending it to Home Assistant
4. Prints the relative URL of the text to speech (TTS) response to the voice command
5. Repeats steps 2-4 until you quit with CTRL+C

Running

arecord -r 16000 -c 1 -f S16_LE -t raw | \
    python3 audio_to_audio.py --rate 16000 --width 2 --channels 1 --token '<HA_LONG_LIVED_ACCESS_TOKEN>'

See python3 audio_to_audio.py --help for more options.