Implementing Continuous Voice Recognition in Flutter Apps

In this tutorial, I will share how to implement continuous voice recognition in a Flutter app. To achieve this, I will use a Flutter package called voice_recognition_flutter. I am excited to announce that I developed this package, so let's dive into the implementation.

install package

dependencies:
  voice_recognition_flutter: ^1.1.0

after install go to the android/app/src/main/AndroidManifest.xml to add following permission

    <uses-permission android:name="android.permission.RECORD_AUDIO" />
    <uses-permission android:name="android.permission.INTERNET" />
    <queries>
        <intent>
            <action android:name="android.speech.RecognitionService" />
        </intent>
    </queries>

after adding these following permission jump to main.dart file to write logic to get a some interesting view

import 'dart:async';

import 'package:flutter/material.dart';
import 'package:voice_recognition_flutter/voice_recognition.dart';
void main() {
  runApp( const MyApp());
}
class MyApp extends StatelessWidget {
  const MyApp({super.key});

  @override
  Widget build(BuildContext context) {
    return const MaterialApp(
      title: 'Voice Recognition Demo',
      home: VoiceRecognitionDemo(),
    );
  }
}

class VoiceRecognitionDemo extends StatefulWidget {
  const VoiceRecognitionDemo({super.key});

  @override
  _VoiceRecognitionDemoState createState() => _VoiceRecognitionDemoState();
}

class _VoiceRecognitionDemoState extends State<VoiceRecognitionDemo> {
  final VoiceRecognition _recognition = VoiceRecognition();
  String _text = '';
  bool _isListening = false;
  double _voiceLevel = 0.0;
  List<String> _locales = [];
  String _currentLocale = "";

  @override
  void initState() {
    super.initState();
    scheduleMicrotask(()async{
     await  _loadLocales();
    });
    _initRecognition();
  }

  Future<void> _loadLocales() async {
    final locales = await _recognition.getAllLocal();
    print(locales.runtimeType);
    setState(() {

      _locales.clear();
      //remove duplicates to avoid dropdown error
      _locales.addAll(locales.toSet().toList());
      _currentLocale = locales.first;

    });
  }


  void _initRecognition() {
    _recognition.setLanguages(_currentLocale);

    _recognition.listenResult(
      onReadyForSpeech: (onReadyForSpeech) {},
        onBeginningOfSpeech: (onBeginningOfSpeech) {},
        onRmsChanged: (rms) {
          setState(() => _voiceLevel = rms! / 10);
        },
        onBufferReceived: (onBufferReceived) {},
        onEndOfSpeech: (onEndOfSpeech) {},
        onError: (onError)async {
        print("Error: $onError");
        await _recognition.stopVoice();
        setState(() => _isListening = false);
        },
        onResults: (onResults) async {
        setState(() => _text += "$onResults ");

        //call this method for Continuous voice recognition  
        await _recognition.startVoice();
        },
        onPartialResults: (onPartialResults) {},
        onEvent: (onEvent) {});

  }

  @override
  Widget build(BuildContext context) {
    print(_locales.length);
    return Scaffold(
      appBar: AppBar(title: const Text('Voice Recognition')),
      body: Padding(
        padding: const EdgeInsets.all(16),
        child: Column(
          children: [

            DropdownButton<String>(
              value: _currentLocale,
              items: _locales.map((locale) {
                return DropdownMenuItem(
                  value: locale,
                  child: Text(locale),
                );
              }).toList(),
              onChanged: (value) {
                setState(() {
                  _currentLocale = value!;
                  _recognition.setLanguages(_currentLocale);
                });
              },
            ),

            const SizedBox(height: 20),


            const Text("Voice Level:"),
            LinearProgressIndicator(
              value: _voiceLevel,
              minHeight: 15,
              backgroundColor: Colors.grey[300],
              valueColor: AlwaysStoppedAnimation<Color>(
                Colors.blue.withOpacity(_isListening ? 1.0 : 0.3)
              ),
            ),

            const SizedBox(height: 20),


            Expanded(
              child: Container(
                padding: const EdgeInsets.all(12),
                decoration: BoxDecoration(
                  border: Border.all(color: Colors.grey),
                  borderRadius: BorderRadius.circular(8),
                ),
                child: SingleChildScrollView(
                  child: Text(_text.isEmpty ? 'Say something...' : _text),
                ),
              ),
            ),

            const SizedBox(height: 20),


            Row(
              mainAxisAlignment: MainAxisAlignment.spaceEvenly,
              children: [
                ElevatedButton(
                  onPressed: _isListening ? null : () async {
                    setState(() => _isListening = true);
                    await _recognition.startVoice();
                  },
                  child: const Text('Start'),
                ),
                ElevatedButton(
                  onPressed: !_isListening ? null : () async {
                    setState(() => _isListening = false);
                    await _recognition.stopVoice();
                  },
                  child: const Text('Stop'),
                ),
                ElevatedButton(
                  onPressed: () => setState(() => _text = ''),
                  child: const Text('Clear'),
                ),
              ],
            ),
          ],
        ),
      ),
    );
  }
}

Note: you can call startVoice() function in onResults listener for continuous voice recognition

I have tested it works fine with en

output

Happy Coding :)