Added #

Core Features

• Unified Platform API - Consistent cross-platform interface
• Typed Exceptions - ModelException, AudioException, TranscriptionException, PermissionException
• Configuration Presets - Quick configs: fast, balanced, accurate, lowMemory, realtime
• Progress Callbacks - Unified progress reporting for downloads and transcription
• Cancellation Tokens - Cancel long-running operations

Transcription Features

• Batch Transcription - Process multiple files with concurrent/sequential modes
• Export Formats - Export to SRT, VTT, JSON, plain text
• Transcription Caching - File-based caching with expiration
• Timestamp Precision - Word-level timestamp estimation
• Translation Improvements - Post-processing for better quality

Advanced Features

• Language Identification - Support for 99 languages
• Custom Model Loading - Load user-provided GGML/GGUF models
• Adaptive Processing - Device capability detection and optimal settings
• Speaker Diarization - Identify different speakers
• Background Transcription - Process audio in background

Enterprise Features

• Queue Management - Priority queue for batch processing
• Large File Handling - Process files >100MB with chunking
• Database Integration - Abstract database interface
• Cloud Storage - Integration with S3, GCS, Firebase Storage
• WebSocket Support - Real-time transcription streaming
• REST API Wrapper - Integrate with remote transcription APIs

Security & Privacy

• Model Security - On-device model verification
• Secure Model Loading - Trusted sources, quarantine invalid models
• Privacy Options - GDPR-compliant processing modes

Production Features

• Telemetry - Usage analytics integration
• Crash Reporting - Error tracking and breadcrumbs
• A/B Testing - Experiment framework with variants
• Feature Flags - Runtime feature toggling
• Auto-Update - Model version management

Developer Experience

• Flutter Widgets - Pre-built UI components
• Plugin Architecture - Extensible third-party integration
• Memory Optimization - Memory monitoring and thread config
• Benchmarking - Performance measurement tools
• Testing Utilities - Mock builders and assertions
• Sample Templates - Ready-to-use app templates
• Firebase Integration - Firestore and Storage utilities

Documentation #

• API reference guide (doc/API_REFERENCE.md)
• Performance optimization guide (doc/PERFORMANCE_GUIDE.md)
• GDPR compliance documentation (doc/GDPR_COMPLIANCE.md)

Added #

• iOS native implementation (beta)
• Enhanced audio handling on Android
• Improved error handling

Fixed #

• Memory leaks on repeated transcriptions
• Crash on invalid audio format

Added #

• Initial release
• Android platform support
• Basic transcription API
• Model download with progress tracking
• Support for Tiny, Base, Small, Medium models
• Language detection and translation to English
• Timestamped segments

Migration Guide: 0.2.x to 0.3.x #

Using Typed Exceptions #

// Before (0.2.x)
try {
  await whisper.transcribe(...);
} catch (e) {
  print('Error: $e');
}

// After (0.3.x)
try {
  await whisper.transcribe(...);
} on ModelException catch (e) {
  print('Model error: ${e.message}');
} on AudioException catch (e) {
  print('Audio error: ${e.message}');
}

Using Configuration Presets #

// Before
final request = TranscribeRequest(audio: audioPath, threads: 4);

// After
final request = TranscriptionPreset.fast.toRequest(audioPath);

Using Export Formats #

// New in 0.3.x
final srt = response.toSRT();
final vtt = response.toVTT();

Compatibility #