Trace the root cause
Diagnostics and wrong-option tags map a missed item back to the outcome and prerequisite that caused the failure.
A question bank stores items. PrepAI operates the repair loop.
PrepAI repairs root causes and schedules memory, not just searchable questions.
Searchable banks are useful for finding practice. PrepAI adds the missing operating layer: diagnose the real weakness, make the student commit before reveal, schedule the repair, and prove the method returns before the exam.
What search alone cannot decide
1
A random searchable question does not know whether the weak point is concentration units, redox setup, logarithms, or the current chapter.
2
A solved question can make a student feel familiar without proving they can retrieve the method tomorrow.
3
A random practice set can over-drill visible symptoms while the prerequisite keeps decaying.
4
A large bank can still leak answers too early, skip misconception tags, and miss board-style timing pressure.
The root-cause repair loop
PrepAI treats a question as evidence. The goal is not to browse more items. The goal is to identify the weak link, repair it, and bring it back at the right interval.
Commit before reveal
The student answers first. The explanation, rubric, and misconception note unlock only after the attempt is recorded.
Score the method
Stepwise rubric scoring shows where marks were earned or lost: formula, substitution, units, reasoning, and final result.
Schedule the memory
Due cards, weak outcomes, and mistake repairs return through spaced retrieval instead of disappearing after one solution.
Repair until retrieval works
A mistake is not cleared when the student reads the answer. It clears when the student successfully retrieves the method again.
Question bank plus operating system
PrepAI still needs a strong reviewed bank. The difference is that the bank feeds an operating system for root-cause diagnosis, spaced repair, rubric scoring, and exam-format proof.
Random searchable question flow
PrepAI operating loop
Random searchable question flow: search by chapter, keyword, or question type.
Compile the exact retrieval set from mastery, due dates, prerequisites, and recent mistakes.
Random searchable question flow: show a solution after the question.
Mask the answer until commit, then reveal formula, substitution, marks, misconception, and next practice.
Random searchable question flow: count how many questions were attempted.
Track whether the student repaired the underlying weakness through successful spaced retrieval.
Random searchable question flow: collect items without proving board-paper transfer.
Use the exam compiler to turn approved questions into timed, official-format sections with answer-key separation.
Evidence students can act on
The dashboard should not say only that a student attempted more questions. It should show the weak outcome, the scheduled repair, and whether the repair survived retrieval.
Outcome-level diagnosis, not just chapter filtering.
Prerequisite graph links visible symptoms to root concepts.
Daily DPP uses due cards, weak outcomes, and stretch topics.
Stepwise rubric scoring shows formula, substitution, unit, and reasoning marks.
Answer masking protects active recall before explanation reveal.
Mistake Lab schedules 24-hour and 72-hour repair checks.
Mock readiness verifies whether repairs transfer into board-style papers.
Search is a tool
Students can still browse topics and practice types, but browsing is not the primary learning loop.
Scheduling is the engine
FSRS-style due dates and mistake repairs decide when a weak concept should come back.
Rubrics protect marks
Stepwise scoring makes the written method visible instead of treating the final answer as the whole performance.
Review gates protect quality
Student-facing practice depends on reviewed content, source coordinates, safe formulas, and rollback controls.
Claim ceiling
This page explains the product strategy. It does not claim the entire bank is approved, that results are guaranteed, or that provider, payment, school, or official partnership states are live.
No claim that every question is approved.
No claim of guaranteed score improvement.
No official CBSE or NCERT partnership claim.
No live revenue, payment, or school adoption claim.
Use questions to repair the system, not pad the count.
Every attempt should update mastery, mistake state, due date, and next practice priority.