How desukyu/handy works
Handy is a cross-platform desktop dictation app: a user triggers recording, speaks, and gets locally transcribed text inserted into the active app. The README documents local VAD, Whisper and Parakeet transcription, shortcut/push-to-talk operation, and paste behavior; the source index supports the backend coordination and post-processing links below. Settings, history, model management, overlays, tray controls, and optional LLM post-processing are distinct product capabilities. Several backend modules were not sampled, so their internal relationships are omitted where the evidence does not establish them.
Main entry points: actions.rs Transcription actions · transcription_coordinator.rs Transcription coordinator · audio.rs Audio manager · App.tsx React app & settings
Components 21
Capture & control
- Shortcut handler
src-tauri/src/shortcut/handler.rs - Transcription coordinator
src-tauri/src/transcription_coordinator.rs - Audio manager
src-tauri/src/managers/audio.rs
Speech processing
- Audio & VAD
src-tauri/src/audio_toolkit - Transcription actions
src-tauri/src/actions.rs - Transcription manager
src-tauri/src/managers/transcription.rs - Model manager
src-tauri/src/managers/model.rs - Transcript history
src-tauri/src/managers/history.rs
Text & extensions
- Text input
src-tauri/src/clipboard.rs - System input
src-tauri/src/input.rs - LLM post-processing
src-tauri/src/llm_client.rs
Desktop app
- Settings store
src-tauri/src/settings.rs - Tauri commands
src-tauri/src/commands - Frontend bindings
src/bindings.ts - React app & settings
src/App.tsx - Settings state
src/stores/settingsStore.ts - Recording overlay
src-tauri/src/overlay.rs - System tray
src-tauri/src/tray.rs
Other components
- User
- Active text app
- Configured LLM API
Connections 23
- User → Shortcut handler: triggers recording
- Shortcut handler → Transcription coordinator: dispatches action
src-tauri/src/shortcut/handler.rs - Transcription coordinator → Audio manager: controls recording
src-tauri/src/transcription_coordinator.rs - Transcription coordinator → Transcription actions: dispatches
src-tauri/src/transcription_coordinator.rs - Audio manager → Audio & VAD: captures audio
- Transcription actions → Transcription manager: transcribes audio
src-tauri/src/actions.rs - Transcription actions → Transcript history: stores transcripts
src-tauri/src/actions.rs - Transcription actions → LLM post-processing: requests post-processing
src-tauri/src/actions.rs - Transcription actions → Audio & VAD: uses audio utilities
src-tauri/src/actions.rs - Transcription manager → Model manager: uses selected model
- User → Active text app: speaks into
- Transcription actions → Text input: delivers transcript
- Text input → System input: types or pastes
src-tauri/src/clipboard.rs - System input → Active text app: inserts text
- LLM post-processing → Configured LLM API: sends optional request
src-tauri/src/llm_client.rs - React app & settings → Frontend bindings: invokes commands
src/bindings.ts - Frontend bindings → Tauri commands: calls backend
src/bindings.ts - Tauri commands → Settings store: reads and updates
src-tauri/src/commands/mod.rs - Settings state → Frontend bindings: loads and updates settings
src/stores/settingsStore.ts - React app & settings → Settings state: uses settings state
src/App.tsx - User → React app & settings: configures app
- User → System tray: controls app
- Audio manager → Recording overlay: updates recording display