supertonic_flutter 2.0.0

High-quality multilingual text-to-speech engine for Flutter. Supports 31 languages with multiple voice styles. Powered by Supertonic 3.

2.0.0

Breaking — upgraded to Supertonic 3.

• Language support expanded from 5 to 31 languages (English, Korean, Japanese, Spanish, Portuguese, French, German, Italian, Dutch, Russian, Ukrainian, Polish, Czech, Slovak, Croatian, Slovenian, Bulgarian, Romanian, Hungarian, Greek, Swedish, Danish, Finnish, Estonian, Latvian, Lithuanian, Turkish, Arabic, Hindi, Indonesian, Vietnamese)
• Model assets now downloaded from Supertone/supertonic-3 (~400 MB)
• Versioned model cache (supertonic_models_v3): existing Supertonic 2 caches are orphaned and the new model is re-downloaded on first run, preventing silent stale-model loads
• Example app: flags added for all 31 languages
• Maintenance: Flutter migrator gradle flags applied

1.0.0

• Auto-download model files from HuggingFace when not bundled as assets
• Three-strategy model loading: local cache → bundled Flutter assets → network download
• Added SupertonicTTS.preDownloadModels() for explicit model pre-downloading with progress callback
• Added SupertonicTTS.modelsReady() to check if models are available locally
• Added DownloadProgressCallback typedef for download progress reporting
• Bundled assets still supported — no code changes needed for existing setups

0.1.5

• Added Web platform support
• Added web plugin registration (SupertonicFlutterWebPlugin)
• Reworked ONNX model loading to use asset-based session creation across platforms
• Added Web setup instructions for loading onnxruntime-web in web/index.html
• Fixed Web TTS inference for tensor(int64) model inputs
• Updated audio playback to in-memory byte source for better cross-platform compatibility

0.1.4

• Added Linux desktop support
• Migrated from just_audio to audioplayers for cross-platform audio playback

0.1.3

• Removed verbose logs at init

0.1.2

• Added comprehensive inline documentation for all public APIs
  • Enhanced TTSConfig with detailed parameter documentation including quality/speed trade-offs
  • Added detailed voice style descriptions for all 10 voices (M1-M5, F1-F5)
  • Documented all supported languages with examples
  • Added usage examples throughout the API
• Removed redundant gender field from TTSVoiceStyle (gender is indicated by voice code prefix)
• Improved code documentation for better IDE hover tooltips
• Updated example app to include silence duration slider, making all TTSConfig parameters adjustable from UI

0.1.1

• Minor README improvements

0.1.0

• Initial release of supertonic_flutter
• Support for multiple languages: English, Korean, Spanish, Portuguese, and French
• Multiple voice styles for each language
• Cross-platform support: Android, iOS, and macOS
• ONNX-based text-to-speech engine
• Configurable speech rate, pitch, and volume
• Audio playback with pause, resume, and stop controls
• Stream-based audio generation