#Setup with Docker?

1 messages · Page 1 of 1 (latest)

vapid wind
#

The setup hasn't been the smoothest so far, mostly because I had to change the home assistant network settings to use a deciated hostname rather than the automatic local ip, and I had to manually enter the IP address of the voice pe to esphome, rather than it being automatic.

I'm now stuck here - I can't actually get whisper or piper to work. I am using a container, and so it can't install the add-ons automatically, but it wants a host and port for the wyoming protocol. It feels like there is a lot of assumed knowledge which I don't have.

How can I get this to work? Is the host supposed to be the address of voice pe? What port is it? It's not clear from any docs :(

whole charm
#

You have to run those services somewhere and then point to them from HA

vapid wind
#

Okay, I thought as much, thanks

opal dust
#

You can use this compose:

https://community.home-assistant.io/t/is-it-possible-to-run-ha-voice-on-docker-or-forget/641454/2

I HA, install the Wyoming Protocol integration and add the endpoints you just created.

hollow raptor
#
services:
  # https://github.com/linuxserver/docker-piper/
  piper:
    restart: unless-stopped
    container_name: piper
    image: lscr.io/linuxserver/piper:latest
    ports:
      - 10200:10200
    env_file:
      - piper.env 
    volumes:
      - /data/piper:/config

  whisper:
    restart: unless-stopped
    image: rhasspy/wyoming-whisper:latest
    container_name: whisper
    ports:
      - 10300:10300
    volumes:
      - '/data/whisper:/data'
    command: '--model small --language en'

  openwakeword:
    image: rhasspy/wyoming-openwakeword:latest
    restart: unless-stopped
    container_name: openwakeword
    ports:
      - '10400:10400'
    volumes:
      - /data/openwakeword:/custom
    command: "--preload-model 'hey_jarvis' --custom-model-dir /custom"
``` is my own relevant compose entry - I use LSIO's Piper as it supports the voice model I want, where the stock one doesn't for some reason