Picovoice Porcupine

Porcupine is a highly-accurate, lightweight, and computationally-efficient wake word engine designed for always-listening voice-enabled applications. It is suitable for IoT devices and supports multiple simultaneous voice commands without increasing the runtime footprint.

Key Features:

• Deep Neural Networks: Trained in real-world environments for high accuracy.
• Cross-Platform Support: Works on Arm Cortex-M, STM32, Arduino, Raspberry Pi, Android, iOS, Web browsers (Chrome, Safari, Firefox, Edge), Linux, macOS, and Windows.
• Customization: Developers can train custom wake word models using the Picovoice Console.
• Scalability: Detects multiple commands with no added runtime footprint.

Replace the placeholder in the lib/main.dart file with your actual AccessKey obtained from the Picovoice Console.

final String accessKey = "{YOUR_ACCESS_KEY_HERE}"; // AccessKey obtained from Picovoice Console (https://console.picovoice.ai/)

To detect keywords within a specific audio file:

Build

cmake -S demo/c/. -B demo/c/build && cmake --build demo/c/build --target porcupine_demo_file

Run (Linux, macOS, Raspberry Pi)

./demo/c/build/porcupine_demo_file -l ${LIBRARY_PATH} -m lib/common/porcupine_params.pv \
-k resources/keyword_files/${PLATFORM}/porcupine_${PLATFORM}.ppn -t 0.5 \
-w resources/audio_samples/multiple_keywords.wav -a ${ACCESS_KEY}

Run (Windows)

.\demo\c\build\porcupine_demo_file.exe ^
-l lib/windows/amd64/libpv_porcupine.dll ^
-m lib/common/porcupine_params.pv ^
-k resources/keyword_files/windows/porcupine_windows.ppn -t 0.5 \
-w resources/audio_samples/multiple_keywords.wav -a ${ACCESS_KEY}

To use Porcupine in a React application, install both the Porcupine React binding and the Web Voice Processor package using npm or yarn.

yarn add @picovoice/porcupine-react @picovoice/web-voice-processor

# or

npm install --save @picovoice/porcupine-react @picovoice/web-voice-processor

iOS provides two types of demos:

BackgroundService Demo

Runs audio recording in the background to detect wake words when the app is not in focus.

1. Open BackgroundService.xcodeproj in Xcode.
2. Replace accessKey in ViewModel.swift with your AccessKey.
3. Run on a simulator or device.

ForegroundApp Demo

Runs detection only when the application is in focus.

1. Open ForegroundApp.xcodeproj in Xcode.
2. Replace accessKey in ViewModel.swift with your AccessKey.
3. Select the desired language scheme from Product > Scheme (e.g., esDemo).
4. Run on a simulator or device.

Vanilla JavaScript and HTML

From the demo/web directory, install dependencies and start the server:

yarn
yarn start ${LANGUAGE}
# OR
npm install
npm run start ${LANGUAGE}

Access the demo at http://localhost:5000.

React

From the demo/react directory, install dependencies and start the server:

yarn
yarn start ${LANGUAGE}
# OR
npm install
npm run start ${LANGUAGE}

Access the demo at http://localhost:3000.

To include Porcupine in your Android project, ensure mavenCentral() is included in your top-level build.gradle file. Then, add the dependency to your app's build.gradle file using the following implementation line:

dependencies {
    // ...
    implementation 'ai.picovoice:porcupine-android:${LATEST_VERSION}'
}

You can install the Porcupine NuGet package using the .NET CLI or via Visual Studio.

Prerequisites:

• For .NET Standard 2.0+: macOS (x86_64), Windows (x86_64).
• For .NET 6.0+: Linux (x86_64), macOS (x86_64, arm64), Windows (x86_64, arm64), and Raspberry Pi (3, 4, or 5).

dotnet add package Porcupine

Install via yarn add @picovoice/porcupine-node. Create a Porcupine instance with built-in keywords or custom keyword paths. Process audio using the .process() method and call .release() to free WebAssembly resources.

const { Porcupine, BuiltinKeyword } = require("@picovoice/porcupine-node");

const accessKey = "${ACCESS_KEY}";

// Using built-in keywords
let handle = new Porcupine(
    accessKey,
    [BuiltinKeyword.GRASSHOPPER, BuiltinKeyword.BUMBLEBEE],
    [0.5, 0.65]
);

// Using custom keyword path
// let handle = new Porcupine(accessKey, ["/path/to/custom/keyword/file"], [0.5]);

while (true) {
  let keywordIndex = handle.process(getNextAudioFrame());
  if (keywordIndex !== -1) {
    // detection event
  }
}

handle.release();