Skip to main content

​
Synopsis

​
What it does

  • Installs voice runtime dependencies (Whisper via whisper.cpp, Piper TTS, ALSA tools)
  • Downloads STT/TTS model files into the voice assets directory
  • Creates the default kiosk holding and thinking WAVs when possible (or run the kiosk cache commands in Retail Kiosk)
On success, prints the voice assets directory path and suggests trying moorcheh-edge voice ask --kiosk-mode. For Bluetooth speakers, pair manually first - see Bluetooth (manual setup).

​
When to run

Once per edge device, after pip install moorcheh-edge and before voice listen, voice ask, or voice serve.