Skip to content

Roadmap

Planned features and future direction for WakeBuilder.

Current Status

WakeBuilder is currently in Beta status (v0.1.x).

Component Status
AST Training Pipeline ✅ Stable
Web Interface ✅ Stable
TTS Integration ✅ Stable
Hard Negative Generation ✅ Stable
ONNX Export ✅ Stable
Docker Deployment 🚧 In Progress
Default Models 🚧 In Progress

Planned Features

Near Term (v0.2.x)

Feature Description Priority
Docker Compose Complete Docker deployment High
Default Models Pre-trained wake words High
Model Comparison Compare multiple models Medium
Threshold Tuning UI Interactive threshold adjustment Medium
Export Improvements Better ONNX compatibility Medium

Medium Term (v0.3.x)

Feature Description Priority
Multi-Language Non-English wake word support High
Alternative Base Models Whisper, Wav2Vec2 options High
Model Quantization Smaller, faster models Medium
Batch Training Train multiple wake words Medium
Training Presets Quick configuration profiles Low

Long Term (v1.0.x)

Feature Description Priority
WakeEngine Integration Seamless deployment High
Mobile Export TensorFlow Lite support High
Cloud-Optional Optional cloud features Medium
Custom Base Models Bring your own embeddings Medium
Streaming Detection Continuous audio monitoring Medium

Version History

v0.1.0 (Current)

  • Initial release
  • AST-based transfer learning
  • Multi-TTS data augmentation
  • Hard negative generation
  • Web interface with training wizard
  • Real-time testing
  • ONNX export

Contributing to Roadmap

Have a feature idea? Help shape WakeBuilder's future:

  1. Check existing issues for similar requests
  2. Open a new issue with the enhancement label
  3. Describe the use case and potential implementation
  4. Join the discussion!

No Promises

This roadmap represents current plans, which may change based on:

  • Community feedback
  • Technical discoveries
  • Resource availability
  • Emerging requirements

Features may be added, removed, or reprioritized as the project evolves.