Bridge layer for the FreakingGenius edge device app.
Decodes incoming WebSocket messages from the tutor, routes them to the correct rendering layer, and reports device capabilities on connect.
Stylus capture layer for the FreakingGenius edge device app.
Converts raw pointer events from the HTML canvas into typed Stroke objects for transmission over the bridge.
Canvas rendering layer for the FreakingGenius edge device app.
Paints exercise content, live student ink, and committed strokes onto the device canvas using the 2D Canvas API.
Device state store for the FreakingGenius edge device app.
Observable singleton that tracks connection status, active session, device capabilities, and runtime counters.
Tutor behavior engine for the FreakingGenius tutor app.
Covers the finite-state machine, intervention scheduling, progressive hint escalation, and persona management.
Bridge client for the FreakingGenius tutor app.
WebSocket client that connects the tutor brain to the edge display device, dispatching render commands and receiving stroke events.
Session lifecycle management for the FreakingGenius tutor app.
Covers the top-level session orchestrator, wall-clock tracking, and durable checkpoint persistence.
Ink-to-math vision pipeline for the FreakingGenius tutor app.
Covers stroke grouping → symbol recognition → expression parsing → work analysis and error pattern detection.
Voice pipeline for the FreakingGenius tutor app.
Covers microphone capture → echo cancellation → VAD → STT → response generation → TTS → audio output.
Tutor annotation overlay for the FreakingGenius edge device app.
Manages tutor ink annotations (highlights, correction marks, instructional drawings) overlaid on top of student work.