flutter_davoice 0.0.1

Flutter plugin for Davoice speech recognition and text-to-speech APIs.

flutter_davoice

Flutter plugin for Davoice speech recognition, text-to-speech, audio playback, license, and native voice framework APIs.

This package is the Flutter equivalent of the React Native bridge in ../TTSNPM/speech/index.ts. The native iOS and Android libraries are vendored inside this plugin so a Flutter app can call the same Davoice STT/TTS APIs from Dart.

Native Libraries

The package expects the native artifacts in these locations:

• iOS framework: ios/Frameworks/DavoiceTTS.xcframework
• iOS static libraries: ios/Libraries/libphonemes.a, ios/Libraries/libucd.a
• iOS phoneme module map/header: ios/Libraries/phonemes/
• Android AAR Maven layout: android/libs/com/davoice/tts/1.0.0/

Run ./update_lib_dirs.sh after replacing the Android AAR or POM. It refreshes the Maven checksum files.

Installation

Add the package to your Flutter app. For local development:

dependencies:
  flutter_davoice:
    path: ../FlutterPubDavoicePrivate

Android

The Davoice Android AAR is stored as a local Maven artifact. Add the plugin's android/libs directory to the host app repositories.

For an app next to this repo, android/build.gradle.kts can include:

allprojects {
    repositories {
        google()
        mavenCentral()
        maven {
            url = uri(rootProject.file("../../FlutterPubDavoicePrivate/android/libs"))
        }
    }
}

The plugin requires Android minSdk 29 or newer.

Add app permissions as needed:

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

iOS

The podspec links the vendored framework and static libraries automatically. Add the host app usage descriptions:

<key>NSMicrophoneUsageDescription</key>
<string>This app uses the microphone for voice features.</string>
<key>NSSpeechRecognitionUsageDescription</key>
<string>This app uses speech recognition to convert speech to text.</string>

The plugin supports iOS 13.0 or newer.

Basic Usage

import 'package:flutter_davoice/flutter_davoice.dart';

final speech = FlutterDavoice();

await speech.setLicense('YOUR_LICENSE_KEY');

await speech.initAll(
  const DavoiceInitAllOptions(
    locale: 'en-US',
    model: 'assets/models/model_ex_ariana_fast.dm',
    timeoutMs: 30000,
  ),
);

await speech.speak(
  'Hello from Davoice.',
  speakerId: 0,
  speed: 0.88,
);

await speech.start(
  'en-US',
  options: const DavoiceStartOptions(extraPartialResults: true),
);

Events

Native callbacks are exposed as Dart streams:

final subscriptions = [
  speech.onSpeechStart.listen((_) {
    // Recognition started.
  }),
  speech.onSpeechPartialResults.listen((event) {
    print(event.value.join(' '));
  }),
  speech.onSpeechResults.listen((event) {
    print(event.value.join(' '));
  }),
  speech.onSpeechError.listen((event) {
    print(event.message);
  }),
  speech.onFinishedSpeaking.listen((_) {
    // TTS or marked audio playback finished.
  }),
];

Available event streams:

• onSpeechStart
• onSpeechRecognized
• onSpeechEnd
• onSpeechError
• onSpeechResults
• onSpeechPartialResults
• onSpeechVolumeChanged
• onNewSpeechWAV
• onFinishedSpeaking

API Surface

Speech recognition:

• initAll
• destroyAll
• initWithoutModel
• destroyWithoutModel
• destroyWihtouModel
• start
• startWithSVOnboardingJson
• stop
• cancel
• isAvailable
• isRecognizing
• pauseSpeechRecognition
• unPauseSpeechRecognition
• pauseMicrophone
• unPauseMicrophone
• hasIOSMicPermissions
• requestIOSMicPermissions
• hasIOSSpeechRecognitionPermissions
• requestIOSSpeechRecognitionPermissions

Text-to-speech and playback:

• initTTS
• speak
• stopSpeaking
• playWav
• playPCM
• playBuffer

License and audio:

• setLicense
• isLicenseValid
• setAECEnabled

Android remote STT helpers:

• initAllRemoteSTT
• initAllRemoteSTTAndTTS

Model Assets

Model paths are passed as strings to the native bridge. In a Flutter app, add the models to the app pubspec.yaml:

flutter:
  assets:
    - assets/models/model_ex_ariana_fast.dm
    - assets/models/model_ex_rich_fast.dm

Then pass the same asset key to initAll or initTTS.

Speaker Verification

Speaker verification enrollment is produced by the companion wake-word package and can be supplied to Davoice. Use a saved enrollment JSON file path when you want STT and TTS flows to be speaker verified:

await speech.initAll(
  DavoiceInitAllOptions(
    locale: 'en-US',
    model: 'assets/models/model_ex_ariana_fast.dm',
    onboardingJsonPath: enrollmentJson,
  ),
);

The standalone example app demonstrates this flow with flutter_wake_word.

Example App

The runnable Flutter example app is intentionally kept outside this package repository because it carries large voice, wake-word, and speaker-verification model assets that may require Git LFS.

Use the standalone example app repo:

https://github.com/frymanofer/Flutter_DaVoice

That app mirrors the React Native demo flow:

• voice model selection
• optional license entry
• speaker verification onboarding
• saved speaker signature reuse
• real-time speaker verification before voice activation
• wake-word detection through flutter_wake_word
• Davoice STT/TTS initialization with speaker-verification enrollment
• manual TTS and speech echo testing
• Full AI Chat with Gemini-style turn taking

Validation

Common validation commands:

flutter analyze
flutter test
cd example && flutter test
cd example && flutter build apk --debug
cd example && flutter build ios --no-codesign

The repo also includes:

./update_library.sh

That refreshes Android checksums and runs the package checks.