Skip to content

🛠️ Installation & Setup

You can use Wyoming in your own Python project, deploy it with Docker, or run it as part of Home Assistant via add-ons. It’s designed to be lightweight and embeddable.

📦 Option 1: From Source (Development)

git clone https://github.com/rhasspy/wyoming
cd wyoming
pip install .

Requirements: - Python 3.8+ - No external dependencies - Optional: aiohttp for HTTP servers

🐳 Option 2: Dockerized Services

Many Wyoming-compatible services offer Docker images:

docker run -p 10200:10200   rhasspy/wyoming-whisper   --uri tcp://0.0.0.0:10200

Examples: - wyoming-whisper (ASR) - wyoming-piper (TTS) - wyoming-openwakeword (Wakeword)

🧠 Option 3: Home Assistant Add-ons

If you're using Home Assistant:

  1. Go to Settings > Add-ons > Add-on Store
  2. Search for:
  3. Piper (TTS)
  4. Whisper (ASR)
  5. OpenWakeWord
  6. Install and configure pipeline in Assist Settings

Wyoming is used as the protocol between Assist and these services.

🧪 Option 4: CLI Testing

Use mic_transcribe.py to stream audio to a Wyoming ASR server:

python scripts/mic_transcribe.py   --uri tcp://localhost:10200   --rate 16000 --width 2 --channels 1

🔌 Choosing a URI

Wyoming URIs describe the connection type:

URI Scheme Description
tcp://host:port Network socket (cross-host)
unix:///path Fast local Unix socket
stdio:// Communicate via stdin/stdout

🧰 Useful Tools

  • wyoming-cli: Minimal Wyoming command-line tool (coming soon)
  • client.py: Example implementation of a TCP client
  • server.py: Base class to build custom Wyoming servers

✅ Test It

After starting a service:

echo '{"type":"ping"}' | nc localhost 10200

You should receive a pong response if the server is active.