cordova-plugin-audioinput

Real-time microphone capture for Cordova and Capacitor with a single package.

Use this plugin when you need low-latency PCM chunks in JavaScript (streaming, VAD, waveform analysis, custom DSP) and when MediaRecorder is too high-level or too delayed.

Why This Plugin Exists

Mobile apps often need raw, continuous microphone frames, not just encoded audio blobs.

This plugin gives you:

• low-latency PCM chunk streaming to JS
• file recording support (WAV)
• one package for Cordova + Capacitor
• Android, iOS, and web implementations

Features

• Real-time PCM streaming (audioData / audioinput events)
• Cordova native bridge now streams binary PCM payloads (ArrayBuffer) for lower bridge overhead
• Optional normalization (-1.0 .. 1.0) for easier JS DSP
• Optional WAV recording via fileUrl
• Microphone permission helpers
• TypeScript definitions for Capacitor
• Cordova Web Audio integration (streamToWebAudio, connect, disconnect)

Platform Support

Platform	Cordova	Capacitor
Android	✅	✅
iOS	✅	✅
Browser	✅	✅

Notes:

• Capacitor Android build config defaults to minSdkVersion 24.
• Capacitor iOS podspec uses deployment target 14.0.
• Web support is intended for development/lightweight browser use-cases.

Installation

Cordova

cordova plugin add cordova-plugin-audioinput

Capacitor

npm install cordova-plugin-audioinput
npx cap sync

iOS Permission String

Ensure NSMicrophoneUsageDescription exists in your app Info.plist.

Example:

<key>NSMicrophoneUsageDescription</key>
<string>This app needs microphone access to capture audio.</string>

Quick Start (Capacitor)

import { AudioInput } from 'cordova-plugin-audioinput';

await AudioInput.initialize({
  sampleRate: 44100,
  bufferSize: 16384,
  channels: 1,
  format: 'PCM_16BIT',
  normalize: true,
});

const permission = await AudioInput.checkMicrophonePermission();
if (!permission.granted) {
  const requested = await AudioInput.getMicrophonePermission();
  if (!requested.granted) throw new Error('Microphone permission denied');
}

const audioDataHandle = await AudioInput.addListener('audioData', event => {
  // event.data is number[]
  console.log('samples:', event.data.length);
});

const errorHandle = await AudioInput.addListener('audioError', event => {
  console.error('audio error:', event.message);
});

await AudioInput.start();

// ... later
await AudioInput.stop();
await audioDataHandle.remove();
await errorHandle.remove();

Capacitor File Recording

import { AudioInput } from 'cordova-plugin-audioinput';

await AudioInput.addListener('audioInputFinished', event => {
  console.log('WAV file:', event.fileUrl);
});

await AudioInput.start({
  sampleRate: 16000,
  channels: 1,
  format: 'PCM_16BIT',
  fileUrl: 'file:///path/to/recording.wav',
});

// stop() resolves when capture stops.
// fileUrl is delivered via audioInputFinished event.
await AudioInput.stop();

Quick Start (Cordova)

function onAudioInput(event) {
  console.log('samples:', event.data.length);
}

function onAudioInputError(event) {
  console.error('audio error:', event.message);
}

window.addEventListener('audioinput', onAudioInput, false);
window.addEventListener('audioinputerror', onAudioInputError, false);

audioinput.checkMicrophonePermission(function (hasPermission) {
  if (hasPermission) {
    startCapture();
    return;
  }

  audioinput.getMicrophonePermission(function (granted) {
    if (granted) startCapture();
  });
});

function startCapture() {
  audioinput.start({
    sampleRate: 44100,
    bufferSize: 16384,
    channels: 1,
    format: audioinput.FORMAT.PCM_16BIT,
    normalize: true,
  });
}

function stopCapture() {
  audioinput.stop(function (fileUrl) {
    if (fileUrl) console.log('Saved file:', fileUrl);
  });
}

API (Capacitor)

Methods

• initialize(options: AudioInputOptions): Promise<void>
• checkMicrophonePermission(): Promise<{ granted: boolean }>
• getMicrophonePermission(): Promise<{ granted: boolean }>
• start(options?: AudioInputOptions): Promise<void>
• stop(): Promise<{ fileUrl?: string }>
• isCapturing(): Promise<{ capturing: boolean }>
• getCfg(): Promise<AudioInputOptions>
• removeAllListeners(): Promise<void>

Events

• audioData → { data: number[], sampleRate?, channels?, format?, timestamp? }
• audioError → { message: string, code?: string }
• audioInputFinished → { fileUrl: string, timestamp?: number }
• stateChange → { state: 'idle' | 'capturing' | 'stopped' | 'error', message?, timestamp? }

API (Cordova)

Methods

• audioinput.initialize(captureCfg, onComplete)
• audioinput.checkMicrophonePermission(callback)
• audioinput.getMicrophonePermission(callback)
• audioinput.start(captureCfg)
• audioinput.stop(onStopped)
• audioinput.isCapturing()
• audioinput.getCfg()
• audioinput.connect(audioNode)
• audioinput.disconnect()
• audioinput.getAudioContext()

Events

• audioinput → { data, sampleRate?, channels?, format?, timestamp? }
• audioinputerror → { message }
• audioinputfinished → { file, timestamp? }
• audioinputstatechange → { state, message?, timestamp? }

Configuration (AudioInputOptions / captureCfg)

Option	Type	Default	Notes
sampleRate	number	44100	Common values: 8000, 16000, 22050, 44100, 48000
bufferSize	number	16384	Power-of-two is recommended
channels	1 | 2	1	Mono or stereo
format	'PCM_16BIT' | 'PCM_8BIT'	'PCM_16BIT'	PCM_16BIT recommended
normalize	boolean	true	Normalize to float range -1..1
normalizationFactor	number	32767.0	Used when normalize=true
audioSourceType	number	0	See source constants
fileUrl	string	undefined	Record to WAV file instead of streaming events

Cordova-only additions:

Option	Type	Default	Notes
streamToWebAudio	boolean	false	Pipe captured audio through Web Audio API
audioContext	AudioContext	auto	Provide your own context
concatenateMaxChunks	number	10	Queue merge tuning

Constants

Capacitor

import { SampleRate, AudioSourceType } from 'cordova-plugin-audioinput';

Cordova

audioinput.SAMPLERATE
audioinput.CHANNELS
audioinput.FORMAT
audioinput.AUDIOSOURCE_TYPE

Performance Tips

• Prefer PCM_16BIT unless you have a hard requirement for PCM_8BIT.
• Start with mono (channels: 1) and sampleRate: 16000 for speech workloads.
• Use larger bufferSize for lower CPU usage and smaller bufferSize for lower latency.
• If you only need files, set fileUrl and skip streaming processing.
• Run npm run bench:js for a quick synthetic JS hot-path benchmark.

Known Limitations

• Device support for sample-rate/channel combinations varies.
• Bluetooth microphone routing behavior varies by OS/device.
• streamToWebAudio is a Cordova API surface (via window.audioinput).
• Web implementation does not persist fileUrl recordings (it emits an audioError warning and continues streaming).

Demo / Test Apps

• app-audioinput-demo
• Local harnesses in test-apps:
  • test-apps/cordova-test-app
  • test-apps/capacitor-test-app

Changelog

See CHANGELOG.md.

Contributing

PRs are welcome. Please keep backward compatibility for existing Cordova integrations.

License

MIT — see LICENSE.