#Rhasspy-speech standalone docker

1 messages · Page 1 of 1 (latest)

lime lodge
#

I'm trying to manually spin up the wyoming-rhasspy-speech add-on docker container but have some issues making it download exposed entities from home assistant.

My current docker-compose.yml:

services:
  rhasspy-speech:
    container_name: rhasspy-speech
    image: "rhasspy/wyoming-rhasspy-speech:1.0.0"
    restart: unless-stopped
    volumes:
      - "./config/models:/models"
      - "./config/training:/training"
    ports:
      - "10300:10300"
      - "8099:8099"

I've downloaded the en_US-zamia model successfully, but just trying to click "download exposed entities" results in No Home Assistant token error message.

I can't find any documentation regarding any environment variables to pass to the container for this.

Anyone have a clue?

#

Managed to answer it myself:

services:
  rhasspy-speech:
    container_name: rhasspy-speech
    image: "rhasspy/wyoming-rhasspy-speech:1.0.0"
    restart: unless-stopped
    volumes:
      - "./config/models:/models"
      - "./config/training:/training"
    ports:
      - "10300:10300"
      - "8099:8099"     
    command: 
      - "--hass-token=your_token_here"
      - "--hass-websocket-uri=ws://home_assistant_url:8123/api/websocket"

Available parameters found in the GH source code

bright hill
#

Is this "Speech" STT supposed to be more performant than Vosk for example ?