Realtime
Streaming ASR
Chunked, low-latency transcription with partial stabilization before final injection.
Features
Features
The full dictation stack: capture, streaming inference, command interpretation, formatting, and secure injection.
Core Engine
Speech In. Structured Text Out.
On-Device Runtime
Inference runs locally by default with optimized model execution paths for Apple Silicon and Windows x64.
Hotkey Capture
Press-to-talk control with precise stop timing, modifier safety, and clean cursor insertion behavior.
Command Layer
Hands-Free Editing Commands
Voice Edits
- Delete last word / phrase
- Undo last insertion
- Replace prior fragment
Formatting Commands
- New line / new paragraph
- Punctuation directives
- Case and emphasis transforms
Context Profiles
- Per-app behavior tuning
- Custom terminology
- Prompt templates by workflow
Control Plane
Built To Operate, Not Demo
Diagnostics
Error timelines, breadcrumbs, and performance histograms for real troubleshooting.
Observability
Stage-level timing spans with p95 tracking for capture, feature extraction, encoder, and decode.
History
Searchable local transcript history with rapid re-insertion and context recovery.
Need the complete platform behavior? Download the latest build and inspect the diagnostics panel live.