A local-first, privacy-preserving learning system that combines the Feynman Technique with spaced repetition (SM-2) to help you learn deeply and remember permanently.
- Learn by teaching: The best way to understand something is to explain it simply
- Forgetting is natural: Spaced repetition fights the forgetting curve systematically
- Privacy first: Your learning data stays on your machine
- AI as coach, not crutch: Optional AI agents challenge your understanding — they never generate content for you
- Capture — Write what you know about a topic (raw notes)
- Simplify — Explain it like you're teaching someone else
- Identify gaps — What's still unclear? What questions remain?
- Refine — Improve your explanation with analogies and examples
- Review — SM-2 algorithm schedules reviews at optimal intervals
- CLI-based workflow (
study new,study review,study today) - 4-step Feynman technique with version tracking
- Concept-based knowledge graph with typed relationships (parent, prerequisite, related)
- SM-2 spaced repetition scheduling
- macOS notifications for daily review reminders
- Optional AI coaching — planned for Phase 2
Phase 1 (MVP) is complete. Core system is functional: data layer, Feynman workflow, SM-2 review scheduling, concept graph, and CLI.
See docs/ for detailed specifications:
- Python 3.11+
- SQLite (local database)
- typer (CLI)
- Anthropic Claude / OpenAI (optional, user-provided API keys)