Wake Word Engine - macOS Quick Start

  • Wake Word Detection
  • Offline Speech Recognition
  • Local Voice Commands
  • Voice Activation
  • Always Listening
  • macOS
  • x86-64

Requirements

  • Git
  • Python 3
  • Command Line Tools for Xcode
  • Homebrew

Cloning the Repository

Clone the repository using

git clone https://github.com/Picovoice/porcupine.git

Installing Dependencies

Navigate to the root of the repository and install common Python dependencies

pip3 install -r requirements.txt

Install PortAudio

brew install portaudio

Install Python dependencies needed for real-time demo

pip3 install -r demo/python/requirements.txt

Runing the Unit Tests

Test the validity of installation by running Python binding's unit tests

python3 binding/python/test_porcupine.py

Running Real-Time Demo

Run the microphone demo application to monitor the microphone stream for the phrase "Porcupine"

python3 demo/python/porcupine_demo_mic --keywords porcupine

The engine starts processing the audio input from the microphone in realtime and outputs to the terminal when it detects utterances of wake-word "Porcupine".

Creating Custom Wake Words

Enterprises who are commercially engaged with Picovoice can create custom wake word models using Picovoice Console.