Adaptive spaced-repetition study scheduler with RL-based policy learning, quiz generation, and Anki deck export
Export the student's question bank as an Anki .apkg file. Returns the written file path.
Generate n multiple-choice questions on a topic using the configured LLM backend (Claude / custom endpoint / HF / mock).
Get a student's belief state: mastered topics, weak areas, quiz count.
List all students known to SmartStudy.
Ask the RL policy what to do next for this student+topic: 'advance', 'reinforce', or 'review'. Decision comes from the learned Q-table, not an LLM.
Record a quiz result (0.0-1.0). Updates the belief state, trains the Q-learning policy, and reschedules the topic under FSRS.
Topics due for review right now (FSRS memory model), plus the full memory-state schedule: recall probability, stability, next due date.
Save generated questions to the student's bank so they can be exported to Anki. questions_json = JSON array from generate_quiz.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 12 | 2026-07-28+ | v2 |