Wyoming protocol server with streaming support for cloud TTS engines (for now, only Google Cloud and OpenAI).
Ask ChatGPT to tell you a long story, and you will hear the response audio almost immediately instead of waiting for the whole pipeline to finish.
Works with Home Assistant Voice Preview Edition (HAVPE) devices.
This project builds on wyoming-piper by Michael Hansen, licensed under MIT.
-
Add the repository and install Wyoming Cloud Streamer from the Add-on store:
-
Configure Wyoming Cloud Streamer addon settings:
- To use Google Cloud TTS, you need a service account json from Google Cloud. Follow these instructions, (you need only text-to-speech). Set the correct path in the addon settings (the default is
/config/SERVICE_ACCOUNT.json) - To use OpenAI, you need an API key, get it here
- To use Google Cloud TTS, you need a service account json from Google Cloud. Follow these instructions, (you need only text-to-speech). Set the correct path in the addon settings (the default is
-
Configure Wyoming Protocol in Home Assistant:
- Go to Settings => Integrations => Add Integration => Wyoming Protocol
- Add ip/hostname (you can use
127.0.0.1) and port (the default one is10200)
-
Add Wyoming Cloud Streamer to Voice assistants:
- Go to Settings => Voice assistants => choose your assistant => Text-to-speech => Cloud TTS Streamer
- Select the desired voice and language, and you are good to go!