Overview
VAD is a cross-platform Voice Activity Detection package for Flutter, enabling seamless integration of voice detection using Silero VAD v4/v5 models. It allows applications to detect when a user is speaking, making it ideal for voice assistants, real-time transcription, and speech processing workflows. The package supports multiple platforms including Android, iOS, Web, macOS, Windows, and Linux, ensuring consistent behavior across devices. With minimal setup via `flutter pub add vad`, developers can easily incorporate reliable voice activity detection into their apps.
Use cases
- Real-time speech detection
- Voice assistant triggers
- Speech-to-text preprocessing
- Noise filtering in audio streams
- Live transcription applications
- Interactive voice interfaces
Key features
- Supports Silero VAD v4/v5 models
- Cross-platform compatibility
- Lightweight and efficient
- Event-driven voice detection
- Easy integration with Flutter
- Low-latency processing
Suitable for
- Developers building voice assistants
- Apps requiring real-time audio analysis
- Speech recognition pipelines
- Multi-platform audio applications
- Projects using ASR or STT systems
Considerations
- Requires native platform setup for optimal performance
- Model size may affect initial load time
- Best suited for continuous audio input
- Limited customization of detection thresholds
- Dependent on Silero model updates
- May need tuning for non-English speech