react-native-speech-to-text-vin

Speech‑to‑text native module for React Native, using:

• iOS: SFSpeechRecognizer (Speech framework) + AVAudioEngine
• Android: SpeechRecognizer + RecognizerIntent

Provides simple start/stop APIs and events for partial and final transcription, with basic silence auto‑stop on both platforms.

Installation

1. Add the package

Using npm:

npm install react-native-speech-to-text-vin

2. iOS setup
From your React Native project:

cd ios
pod install
cd ..
iOS permissions
In your app’s Info.plist, add:

<key>NSSpeechRecognitionUsageDescription</key>
<string>We need speech recognition to convert your voice to text.</string>
<key>NSMicrophoneUsageDescription</key>
<string>We need access to your microphone for speech recognition.</string>
Without these, iOS will crash or deny access when starting recognition.

3. Android setup
Usually, autolinking is enough. Just rebuild your app:

npx react-native run-android
Android permissions
In android/app/src/main/AndroidManifest.xml ensure you have:

<uses-permission android:name="android.permission.RECORD_AUDIO" />
The JS API will ask for the runtime microphone permission using PermissionsAndroid.

Usage
Basic example (hooks)

import React, { useEffect, useState } from 'react';
import { View, Text, Button } from 'react-native';
import {
  requestAuthorization,
  start,
  stop,
  addResultListener,
  addPartialResultListener,
  addErrorListener,
} from 'react-native-speech-to-text-vin';
export default function SpeechExample() {
  const [partial, setPartial] = useState('');
  const [finalText, setFinalText] = useState('');
  const [status, setStatus] = useState<'idle' | 'listening'>('idle');
  useEffect(() => {
    let resultSub: any;
    let partialSub: any;
    let errorSub: any;
    (async () => {
      const authStatus = await requestAuthorization();
      console.log('Speech auth status:', authStatus);
      resultSub = addResultListener(event => {
        // { value: string }
        setFinalText(event.value);
        setPartial('');
        setStatus('idle');
      });
      partialSub = addPartialResultListener(event => {
        // { value: string }
        setPartial(event.value);
        setStatus('listening');
      });
      errorSub = addErrorListener(event => {
        console.warn('Speech error:', event.error);
        setStatus('idle');
      });
    })();
    return () => {
      stop();
      resultSub?.remove();
      partialSub?.remove();
      errorSub?.remove();
    };
  }, []);
  const handleStart = async () => {
    try {
      await start();
      setStatus('listening');
    } catch (e) {
      console.warn('Failed to start speech recognition', e);
    }
  };
  const handleStop = async () => {
    try {
      await stop();
      setStatus('idle');
    } catch (e) {
      console.warn('Failed to stop speech recognition', e);
    }
  };
  return (
    <View style={{ padding: 16 }}>
      <Text>Status: {status}</Text>
      <Text style={{ marginTop: 16 }}>Partial:</Text>
      <Text>{partial}</Text>
      <Text style={{ marginTop: 16 }}>Final:</Text>
      <Text>{finalText}</Text>
      <View style={{ flexDirection: 'row', marginTop: 24, gap: 12 }}>
        <Button title="Start" onPress={handleStart} />
        <Button title="Stop" onPress={handleStop} />
      </View>
    </View>
  );
}

API
All functions are imported from 'react-native-speech-to-text-vin'.

requestAuthorization(): Promise<string>
Requests speech/microphone authorization.

iOS: Uses SFSpeechRecognizer.requestAuthorization.
Returns one of: "authorized" | "denied" | "restricted" | "notDetermined" | "unknown".
Android: Uses PermissionsAndroid.request for RECORD_AUDIO.
Returns "authorized" if granted, otherwise "denied".
Example:

const status = await requestAuthorization();
if (status !== 'authorized') {
  // Show message / handle denied
}
start(): Promise<void>
Starts speech recognition.

On iOS, starts AVAudioEngine and SFSpeechRecognizer.
On Android, checks RECORD_AUDIO permission, verifies availability, and calls SpeechRecognizer.startListening.
Auto‑stops after a period of silence (implemented natively on both platforms).
Rejects with an error if:

Permission is missing (Android), or
Speech recognition is not available, or
Native setup fails.
stop(): Promise<void>
Stops recognition if running and cleans up internal state.

You should also call this in cleanup (e.g. useEffect return).

addResultListener(listener)
Listens for final recognition results.

Event payload: { value: string }
Emitted when the native recognizer thinks the utterance is complete (or after silence timeout).
Returns a subscription with .remove().

Example:

const sub = addResultListener(event => {
  console.log('Final:', event.value);
});
// later
sub.remove();
addPartialResultListener(listener)
Listens for partial / intermediate recognition results.

Event payload: { value: string }
Can fire multiple times as the user speaks.
addErrorListener(listener)
Listens for recognition errors.

Payload: { error: string }
Platform behaviour details
iOS
Uses:
SFSpeechRecognizer with shouldReportPartialResults = true
AVAudioEngine input tap
Silence handling:
A 2‑second timer resets whenever results come in.
If no new callbacks for ~2 seconds:
If there is a lastTranscript, onSpeechResults is emitted with it.
Recording is stopped.
Android
Uses:
SpeechRecognizer + RecognizerIntent.ACTION_RECOGNIZE_SPEECH
EXTRA_PARTIAL_RESULTS enabled
Silence handling:
A 4‑second timeout resets on partials.
When it fires:
If there is a lastPartialText and no final has been sent, onSpeechResults is emitted.
SpeechRecognizer.stopListening() is called.
Troubleshooting
Native module not linked / not found

If you see:

SpeechToText native module not linked

then autolinking might have failed. Try:

Reinstall node modules
cd ios && pod install
Clean build and rebuild iOS / Android
iOS build errors about Speech/AVFoundation

Make sure you have:

import Speech
import AVFoundation
(already provided in this library’s Swift code)
and that your iOS deployment target in the app is compatible with the podspec (≥ iOS 11.0).

Android permission denied

Ensure both:

RECORD_AUDIO is in AndroidManifest.xml
You are calling requestAuthorization() before start() and handling "denied".