@p0llen/wakeword-react

Status: 🚧 In Testing License: MIT

Free, open-source wake-word detection package using PyTorch, with an associated npm package for React-based integration.
Includes CLI tools, real-time microphone access, and GPU-accelerated training capabilities.

🔧 How It Works

Install the Python backend (wakeword-detector) for model training and ONNX export. Then, install this React package for in-browser wakeword detection.

When sound input matches the trained model, the React component will be notified — enabling real-time browser-based wakeword detection.

🚀 Usage

import { useWakeword, WakewordDemo } from "@p0llen/wakeword-react";

function App() {
  return <WakewordDemo />;
}

🔥 Features

• 🎤 Real-time wakeword detection in the browser
• ⚛️ React component (useWakeword) for easy integration
• 🧠 ONNX model inference with onnxruntime-web
• 🎛️ Adjustable confidence thresholds and visual feedback
• 🔁 Fully customizable — bring your own wakeword
• 🧪 Demo UI (WakewordDemo component) included for instant testing

📦 Installation

1. Backend (Python)

Install the backend model trainer and exporter:

pip install -i https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple wakeword-detector

📋 Requirements

• eact 17 or 18+
• Modern browser (Chrome, Edge, Brave, etc.)
• ONNX model trained using wakeword-detector

🤝 Contributing Found a bug or want to suggest a feature? Open an issue or PR. Contributions welcome!