React Native Audio Recorder

@choiminseok/react-native-audio-recorder is a React Native module that records microphone input, emits real-time PCM chunks, and persists the final WAV file to cache local storage. It targets low-latency speech-to-text and streaming scenarios while remaining simple to integrate in any app.

Basic Functionality

Audio recording

• Call AudioRecorder.startRecording(config?) to capture microphone input using the desired sample rate, channels, chunk size, and Android audio source.
• Pause/resume logic is managed natively to keep latency under 100 ms per chunk while getState() exposes the current RecorderState.
• getDuration() returns the running duration in milliseconds while a session is active.

Real-time audio listener

• Subscribe to audioData via AudioRecorder.addListener('audioData', handler) to receive PCM chunks as Int16 arrays alongside a timestamp and sequence.
• Use these callbacks to feed ASR models, sockets, or visualizers; remove the listener (or call subscription.remove()) when you no longer need streaming data.
• Additional events include stateChange (lifecycle transitions) and error (surface native failures with descriptive codes).

Saving the temporary audio file

• AudioRecorder.stopRecording() flushes all buffered audio, writes a WAV file into the platform cache directory, and resolves with a RecordingResult.
• The result exposes tmpFileUri, durationMs, fileSizeBytes, sampleRate, and channels. You can upload the file immediately or move it to permanent storage.
• Call AudioRecorder.cancelRecording() if you want to discard the session without persisting audio.

Installation

yarn add @choiminseok/react-native-audio-recorder
# or
npm install @choiminseok/react-native-audio-recorder

iOS

cd ios && pod install && cd ..

How to Use

1. Check/request microphone permission using the provided helpers.
2. Attach listeners for audioData, stateChange, or error.
3. Start recording with the desired configuration (or rely on defaults: 16 kHz, mono, 1024-sample chunks).
4. Stop or cancel depending on whether you need the WAV output.
5. Clean up by removing listeners when the component unmounts.

import { useEffect, useState } from 'react';
import AudioRecorder, {
  AudioDataEvent,
  PermissionStatus,
  RecorderState,
} from '@choiminseok/react-native-audio-recorder';

export function UseRecorderExample() {
  const [state, setState] = useState<RecorderState>('idle');
  const [lastChunkTs, setLastChunkTs] = useState<number | null>(null);
  const [tmpFileUri, setTmpFileUri] = useState<string | null>(null);

  useEffect(() => {
    const dataSub = AudioRecorder.addListener('audioData', (event: AudioDataEvent) => {
      setLastChunkTs(event.chunk.timestampMs);
      // sendInt16ToServer(event.chunk.data);
    });
    const stateSub = AudioRecorder.addListener('stateChange', (event) => {
      setState(event.newState);
    });
    return () => {
      dataSub.remove();
      stateSub.remove();
    };
  }, []);

  const ensurePermission = async () => {
    const status = await AudioRecorder.checkPermission();
    if (status !== PermissionStatus.GRANTED) {
      const next = await AudioRecorder.requestPermission();
      if (next !== PermissionStatus.GRANTED) {
        throw new Error('Microphone permission is required');
      }
    }
  };

  const start = async () => {
    await ensurePermission();
    await AudioRecorder.startRecording({
      sampleRate: 16000,
      channels: 1,
      chunkSize: 1024,
    });
  };

  const stop = async () => {
    const result = await AudioRecorder.stopRecording();
    setTmpFileUri(result.tmpFileUri);
  };

  return null;
}

API Surface

| Method | Description | | --- | --- | | checkPermission(): Promise<PermissionStatus> | Returns current microphone authorization state. | | requestPermission(): Promise<PermissionStatus> | Prompts the user if permission is not granted. | | startRecording(config?: RecorderConfig): Promise<void> | Begins recording using the provided configuration. | | stopRecording(): Promise<RecordingResult> | Stops recording, writes WAV, and resolves with file metadata. | | cancelRecording(): Promise<void> | Stops recording without writing a file. | | getState(): Promise<RecorderState> | Returns native recorder state (idle, recording, etc.). | | getDuration(): Promise<number> | Current session duration in milliseconds. | | addListener(event, handler) / removeListener(event) / removeAllListeners(event) | Manage audioData, stateChange, error subscriptions. |

RecorderConfig

• sampleRate (8000 | 16000 | 44100 | 48000, default 16000)
• channels (1 mono or 2 stereo, default 1)
• chunkSize (samples per emitted chunk, default 1024)
• audioSource (Android: default | mic | voiceRecognition | voiceCommunication, default voiceRecognition)
• outputPath (optional absolute destination; otherwise a temp file is created)

Events

• audioData: { chunk: { data: number[]; timestampMs; sequenceNumber } }
• stateChange: { oldState, newState }
• error: { error: { code, message, nativeError? } }

Example App

Run the bundled example to validate native integrations or test changes:

yarn
yarn example:ios   # or yarn example:android

The example already demonstrates streaming chunks, live duration, and file saving flows.

Repository Structure

react-native-audio-recorder/
├── packages/react-native-audio-recorder   # Library source + native code
└── apps/example                           # Expo-style test bed (RN 0.81.5)

License

MIT © CHOIMINSEOK