About this project
Principia Desk is a desktop study application designed for deep, foundational learning. It provides nine built-in classes (Linux Bash, Bash Scripting, JavaScript & Browser, TypeScript, Frontend Architecture, Developer Tooling, System Design, German, and Italian) and allows users to create their own custom classes. The app emphasizes personal starting points, lessons written by an AI tutor from verified primary documentation, durable appointments, and user-chosen enforcement policies.
Key features include:
- **Today view**: Combines class appointments, showing due study times and spaced-review topics.
- **Classes**: Each class has its own tutor, study times, enforcement policy, personal path, and curriculum map.
- **Study times**: Recurring rules with configurable start times, weekdays, and lengths (10 minutes to 8 hours).
- **Starting points**: Choose foundations, declare a starting stage, or take a placement check that samples every stage.
- **Personal paths**: Revise your route with statuses like Completed, Prior knowledge checked, Bypassed by choice, Needs refresher, Not assessed, Bridge lesson, In progress, Upcoming.
- **Custom classes**: A five-step builder (Brief, Draft, Review, Verify, Publish) lets you create courses with AI assistance or from scratch, including question banks.
- **Shared study runtime**: Lessons are planned, prepared, published as immutable versions, and finished in one transaction. Progress is saved as you go.
- **Listening**: Lessons can be converted to audio conversations with selectable voices (system, Piper, Kokoro, VibeVoice).
- **Enforcement**: Per-class focus policies (advisory, focused, strict) with multiple escape hatches: finish the check, break-glass phrase, recovery console (Ctrl+Alt+Shift+U), release token, and a dead man's switch.
- **Content generation**: Supports external agents (Claude Code, Codex, etc.), API keys (Anthropic, OpenAI, etc.), or local models via Ollama.
- **Sourcing**: Fetches primary documents from an allowlist of publishers (MDN, web.dev, W3C, etc.) for lesson generation.
- **Always-on**: Menu bar desk with scheduler, notifications, and study alarms.
Built with Tauri, Svelte, and Rust, it is MIT-licensed and available for macOS, Windows, and Linux.
Comments
0 people shared their preference · Deer Point appears after 10 participants
Sign in to join the discussion.