react-native-nitro-speech-recognition

A powerful speech recognition library for React Native, built with Nitro Modules.

Features

• 🚀 Fast & Efficient: Built with Nitro Modules for high performance.
• 🎙️ Real-time Recognition: Supports real-time speech-to-text.
• 📱 On-Device Support: Supports on-device speech recognition (Android 13+).
• 🌍 Multi-language: Supports multiple locales.
• 🔄 Flexible Audio Input: Accepts raw audio buffers (PCM Int16 or Float32).

Installation

npm install react-native-nitro-speech-recognition
# or
yarn add react-native-nitro-speech-recognition
# or
bun add react-native-nitro-speech-recognition

Expo

You can use this library with Expo. It includes a Config Plugin to automatically configure permissions.

First, install the package:

npx expo install react-native-nitro-speech-recognition

Then, add the config plugin to your app.json or app.config.js:

{
  "expo": {
    "plugins": [
      [
        "react-native-nitro-speech-recognition",
        {
          "microphonePermission": "Allow $(PRODUCT_NAME) to access your microphone.",
          "speechRecognitionPermission": "Allow $(PRODUCT_NAME) to access speech recognition."
        }
      ]
    ]
  }
}

Manual Installation (Bare React Native)

If you are not using Expo, or if you are using a bare React Native project, you need to configure permissions manually.

iOS

Add the following key to your Info.plist to request speech recognition permission:

<key>NSSpeechRecognitionUsageDescription</key>
<string>Allow access to speech recognition for transcribing your voice.</string>
<key>NSMicrophoneUsageDescription</key>
<string>Allow access to the microphone for recording audio.</string>

Android

Add the following permission to your AndroidManifest.xml:

<uses-permission android:name="android.permission.RECORD_AUDIO" />
<uses-permission android:name="android.permission.INTERNET" />

Usage

Basic Usage

import { SpeechRecognition } from 'react-native-nitro-speech-recognition';

const speechRecognition = new SpeechRecognition();

// 1. Check availability
const isAvailable = SpeechRecognition.isRecognitionAvailable();

// 2. Request permissions
const permissions = await SpeechRecognition.requestPermissionsAsync();

// 3. Setup event listeners
speechRecognition.on('result', (event) => {
  console.log('Transcript:', event.results[0]?.transcript);
  console.log('Is Final:', event.isFinal);
});

speechRecognition.on('error', (error) => {
  console.error('Error:', error);
});

// 4. Start recognition
speechRecognition.start({
  locale: 'en-US',
  interimResults: true,
  audioFormat: 'pcmFloat32', // or 'pcmInt16'
  sampleRate: 16000,
});

// 5. Feed audio data (e.g., from a microphone stream)
// This library does NOT handle audio recording. You need to use an external library
// like `react-native-audio-api` to record audio and feed the raw PCM data (ArrayBuffer)
// to the recognizer.

// Example using `react-native-audio-api`:
// import { AudioRecorder } from 'react-native-audio-api';

// const recorder = new AudioRecorder({
//   sampleRate: 16000,
//   bufferLengthInSamples: 4096,
// });

// recorder.onAudioReady((event) => {
//   const bufferData = event.buffer.getChannelData(0);
//   // Ensure the data is in the correct format (e.g. pcmFloat32 or pcmInt16)
//   speechRecognition.streamInsert(bufferData.buffer);
// });

// recorder.start();

// 6. Stop recognition
speechRecognition.stop();

On-Device Recognition (Android)

To use on-device recognition on Android, you might need to download the model first.

const locale = 'en-US';
const supported = await SpeechRecognition.getSupportedLocales();

if (!supported.installedLocales.includes(locale)) {
  const result = await SpeechRecognition.downloadOnDeviceModel(locale, (progress) => {
    console.log(`Download progress: ${progress * 100}%`);
  });
  
  if (result.status === 'download_success') {
    console.log('Model downloaded successfully');
  }
}

speechRecognition.start({
  locale,
  requiresOnDeviceRecognition: true,
  // ... other options
});

API Reference

SpeechRecognition

The main class for managing speech recognition sessions.

Methods

• start(options: SpeechRecognitionOptions): void: Starts the speech recognition session.
• stop(): void: Stops the speech recognition session.
• streamInsert(buffer: ArrayBuffer): void: Feeds audio data into the recognizer.
• on(event: EventType, listener: Function): void: Registers an event listener.
• unsubscribeAll(): void: Unsubscribes all listeners.

Static Methods

• isRecognitionAvailable(): boolean: Checks if speech recognition is available on the device.
• isOnDeviceRecognitionAvailable(): boolean: Checks if on-device recognition is available.
• downloadOnDeviceModel(locale: string, onProgress: (progress: number) => void): Promise<OnDeviceModelDownloadResult>: Downloads the on-device model for the specified locale (Android).
• getSupportedLocales(): Promise<SupportedLocales>: Returns the list of supported and installed locales.
• getPermissionsAsync(): Promise<PermissionResponse>: Gets the current permission status.
• requestPermissionsAsync(): Promise<PermissionResponse>: Requests speech recognition permissions.

SpeechRecognitionOptions

| Property | Type | Description | | :--- | :--- | :--- | | locale | string | The locale for recognition (e.g., "en-US"). | | interimResults | boolean | Whether to return interim results. | | maxAlternatives | number | Maximum number of alternative transcripts. | | requiresOnDeviceRecognition | boolean | Whether to require on-device recognition. | | addsPunctuation | boolean | Whether to add punctuation to the transcript. | | audioFormat | 'pcmInt16' \| 'pcmFloat32' | The format of the input audio buffer. | | sampleRate | number | The sample rate of the input audio. |

Contributing

See the contributing guide to learn how to contribute to the repository and the development workflow.

License

MIT