Wyoming macOS TTS

Wyoming protocol server for Text-to-Speech on macOS, using the say built-in command to generate speech natively on Apple silicon for voice pipelines in Home Assistant.

Features

• Single install command to run in the background and automatically on login
• Supports multiple languages and voices
• Supports streaming

For Speech-to-Text on macOS, check out wyoming-macos-stt

Getting Started

Requirements

• macOS 26
• uv
• ffmpeg

These can be installed using Homebrew - brew install uv ffmpeg

Installing

1. Clone this repository and navigate into it:

   git clone https://github.com/openhoster/wyoming-macos-tts
   cd wyoming-macos-tts

2. Run the installer and follow the prompts:

   uv run script/install.py

   This will create a launcher file and optionally set it to run in the background and on login.

   By default the server will be available externally for devices on the local network. You can change this and other arguments by editing the launcher file WyomingTTS.

   Streaming is disabled by default, to enable it add --streaming in the launcher file.

   To see all the available arguments, run: uv run -m wyoming_macos_tts --help

3. Adding to Home Assistant:
   Add a new Wyoming service with the Mac’s host/IP and port 10200.

Tip: Additional voices can be downloaded in the system settings under Accessibility -> Live Speech -> Voice -> ℹ️.

if new voices are not showing in Home Assistant, try reloading the Wyoming service in Home Assistant

Uninstalling

Re-run the installer and, when prompted: Run in the background and on login? enter n

To completely remove all resources, delete the cloned repository folder.

Development

• Install dev dependencies:

  uv sync --extra dev

• Running in a terminal session:

  uv run -m wyoming_macos_tts --help

• Running tests:

  uv run -m pytest tests

---

If you find this usefull and want to support future projects: