@reactxnative/react-native-voice

🎥 Click here to watch the Demo Video!

A high-performance, custom React Native voice recognition library built entirely using the new Android Native TurboModule Architecture (Bridgeless-ready).

It uses Android's built-in SpeechRecognizer API to deliver fast, offline-capable (depending on OS), and highly accurate voice-to-text recognition with no external dependencies. Currently optimized for Android.

Features

• 🚀 TurboModule Architecture: Blazing fast performance using the latest React Native architecture.
• 🔗 Bridgeless Compatible: Fully supports React Native 0.74+ and 0.86+ New Architecture defaults.
• 🎤 Zero Dependencies: Uses pure Android native SpeechRecognizer class.
• ⚡ Auto-Unwrapping: Super clean JS API that returns raw strings instead of bulky objects.
• 🇮🇳 Multi-Lingual: Supports free-form language modeling (default initialized for Hindi/English but highly extensible).

Installation

npm install @reactxnative/react-native-voice

or

yarn add @reactxnative/react-native-voice

Android Permissions Setup

Since this library interacts with the device's hardware microphone, you must add the following permissions to your android/app/src/main/AndroidManifest.xml:

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

(Note: You must also request runtime permissions in your JS code using PermissionsAndroid.request before calling startListening()).

Usage Example

import { useEffect, useState } from 'react';
import { Button, Text, View } from 'react-native';
import { 
  startListening, 
  stopListening, 
  destroy, 
  addSpeechResultsListener, 
  addSpeechErrorListener 
} from '@reactxnative/react-native-voice';

export default function App() {
  const [text, setText] = useState('');
  const [error, setError] = useState<number | null>(null);

  useEffect(() => {
    // 1. Subscribe to Speech Results
    const resultsSub = addSpeechResultsListener((spokenText: string) => {
      console.log('User said:', spokenText);
      setText(spokenText);
    });

    // 2. Subscribe to Speech Errors
    const errorSub = addSpeechErrorListener((errCode: number) => {
      console.error('Speech Error Code:', errCode);
      setError(errCode);
      
      // Common error codes: 
      // 6 = Timeout (No speech heard)
      // 7 = No match (Speech not recognized)
    });

    return () => {
      // Clean up listeners and destroy native recognizer when component unmounts
      resultsSub.remove();
      errorSub.remove();
      destroy();
    };
  }, []);

  return (
    <View style={{ flex: 1, justifyContent: 'center', alignItems: 'center' }}>
      <Text style={{ fontSize: 24 }}>{text || "Press mic to speak..."}</Text>
      
      <Button title="Start Listening" onPress={() => {
        setError(null);
        startListening();
      }} />
      
      <Button title="Stop" onPress={() => stopListening()} />
      
      {error && <Text style={{ color: 'red' }}>Error Code: {error}</Text>}
    </View>
  );
}

API Reference

Methods

| Method | Description | |--------|-------------| | startListening() | Initializes the native SpeechRecognizer (if not already created) and begins listening to user audio. | | stopListening() | Instructs the recognizer to stop listening. It may still process audio already captured. | | destroy() | Completely destroys the native recognizer instance and frees up memory. Call this when your component unmounts. |

Event Listeners

| Method | Callback Type | Description | |--------|--------------|-------------| | addSpeechResultsListener | (spokenText: string) => void | Fires when the recognizer successfully determines a final result. Returns a clean string. | | addSpeechErrorListener | (errorCode: number) => void | Fires when a native Android speech error occurs. |

Common Android Speech Error Codes

• 1 - Network operation timed out.
• 2 - Other network related errors.
• 3 - Audio recording error.
• 4 - Server sends error status.
• 5 - Other client side errors (e.g., cancelled by user).
• 6 - No speech input (User was silent for 10 seconds).
• 7 - No recognition result matched (Couldn't understand the speech).
• 8 - RecognitionService busy.
• 9 - Insufficient permissions (You forgot to request RECORD_AUDIO).

Contributing

Pull requests are welcome! For major changes, please open an issue first to discuss what you would like to change.

License

MIT