Wyoming Azure Speech

This project implements a Wyoming protocol server backed by Microsoft Azure's AI
Speech services for speech-to-text transcription and text-to-speech
synthesization.

It includes both a Python package and a pre-built Docker image for simpler
integration.

Prerequisites

In order to use the Azure AI Speech Service, one must sign up for a (free)
account on Azure at https://portal.azure.com/. As of the writing of this
document, the free tier offers 5 hours for speech-to-text conversion per month
and 0.5 million characters of text-to-speech. Only when those are exceeded, you
will be billed. For more information, check out the Speech Services
pricing.

Before running the Wyoming server, you need to provide it with the Azure region
serving the Speech Service as well as a subscription key. Those can be generated
by performing the following:

Create an account at Azure
Add a new subscription
Create a new Speech service
- Create a new resource group
- Choose the region closest to you
- Choose the "Free F0" pricing tier
Make a note of one of the generated keys and the region name for later

Installation

Running the server

Manual

Install and the package:

pip install wyoming-azure-speech
wyoming-azure-speech --key <KEY> --region <REGION>

Docker

Pull image and run:

docker run ghcr.io/shmuelzon/wyoming-azure-speech:latest --key <KEY> --region <REGION>

Docker Compose

services:
  wyoming-azure-speech:
    container_name: wyoming-azure-speech
    image: ghcr.io/shmuelzon/wyoming-azure-speech:latest
    restart: unless-stopped
    pull_policy: always
    environment:
      AZURE_KEY_FILE: /run/secrets/azure_key
      AZURE_REGION_FILE: /run/secrets/azure_region
    secrets:
      - azure_key
      - azure_region

secrets:
  azure_key:
    file: secrets/azure_key.txt
  azure_region:
    file: secrets/azure_region.txt

Add to Home Assistant

You can either click on the image or follow the manual procedure below:

Browse to your Home Assistant instance
Go to Settings > Devices & Services
In the bottom right corner, select the + Add Integration button
From the list, select Wyoming Protocol.
Enter the IP and port that the Wyoming server is listening on

Configuration

The server allows the following configuration options:
| Name | Description | Default Value | Optional environment variable |
| --- | --- | :-: | --- |
| key | Azure subscription key | | AZURE_KEY or AZURE_KEY_FILE to read it from a file |
| region | Azure region name | | AZURE_REGION or AZURE_REGION_FILE to read it from a file |
| uri | A URI the server will listen on | tcp://0.0.0.0:10300 | |
| transcription-language | Default language for transcription | en-US | |
| voice | Default voice for synthesizing | en-US-AvaMultilingualNeural | |
| debug | Enable debug logs | False | |
| log-format | The format used for printing log messages | %(asctime)s:%(levelname)s:%(name)s:%(message)s | |

README