The friction before the build.
Cloud dictation services send sensitive audio to external servers, require internet connectivity, introduce latency and reliability concerns, and may retain or analyze recordings.
Productivity In Development ~95% Complete
Toko
Privacy-first voice dictation for macOS.
Overview
What this file is about.
Toko provides fast, accurate voice-to-text without sending audio to the cloud. Using WhisperKit for on-device transcription, it delivers professional-grade dictation with complete privacy—ideal for legal, medical, and security-conscious professionals.
What changed.
- ✓ On-Device Transcription — WhisperKit (CoreML) processes audio locally
- ✓ Global Hotkey — Press and speak from any application
- ✓ Grammar Correction — Automatic cleanup of dictation artifacts
- ✓ Transcription History — Searchable archive with favorites
- ✓ Clipboard Integration — Results ready to paste immediately
- ✓ Multi-Microphone Support — Select from available audio inputs
Market opportunity
- Voice recognition market exceeds $12B
- Legal, medical, and enterprise segments have strict privacy requirements
- On-device AI now viable on consumer hardware
Traction
- Full application implemented (2,700+ lines of Swift)
- MVVM architecture with clean separation
- Ready for final build and testing
Business model
One-time purchase for privacy-conscious professionals. No subscription, no cloud costs, no data collection.
Interested in Toko?