How Rank Sarthi Works: From Attempt Evidence to the Next Priority

Start from something observable: a mock, practice set, solved paper, revision check or another reviewed attempt record. Preserve question identity, answer state and the student's reasoning evidence where available.

PI v1.1 uses a manual control taxonomy:

• Knowledge Gap: the required concept, fact, formula, relationship or method is not sufficiently known.
• Recall Gap: the learner has learned the material but cannot reliably retrieve it when needed.
• Execution Error: the intended method is appropriate but arithmetic, algebra, units, signs, transcription, reaction handling or procedural execution fails.
• Decision or Selection Error: the learner chooses the wrong method, question order, level of effort, assumption or attempt decision.
• Needs Review: evidence is insufficient or ambiguous, so the error is not forced into another category.

Do not treat every error equally. Prioritise by repeatability, prerequisite impact, severity, time-to-fix and the student's next exam or revision need.

Apply the repair, then test the same underlying concept or decision pattern again. Improvement is evidence of changed performance, not completion of a checklist.

Pick JEE Main, JEE Advanced, NEET UG or NDA. BITSAT, IISER Aptitude Test and NEST are coming soon and cannot be selected yet.

A four-step goal wizard records your target exams and class, a main goal (rank, percentile or score), any extra goals, and a self-assessment of strengths, weak areas and study hours.

One free diagnostic test is available per target exam.

The result shows weak areas and suggests the plan that fits you.

Pick full papers, part papers, previous-year papers or topic practice, filtered by exam, year, subject, topic and subtopic. The Self-Test Builder creates a paper from your chosen subjects, topics and difficulty mix, with exam-pattern presets and a "Practice for my goal" option that picks your weak topics.

Use an NTA-style exam screen with instructions, a question palette, mark-for-review, a timer and section tabs. Answers save automatically, and marking follows each exam's rules, including partial marks where the exam uses them.

Each attempt produces a report with score, rank, percentile, subject-wise and difficulty-wise accuracy, time per question, mistake analysis and strategy tips, plus a downloadable PDF. AI-generated written feedback and an accuracy heatmap by question nature and difficulty explain what went wrong.

The performance dashboard tracks trends over time, weak topics with reasons, comparison with other students and the gap to your goal, so the next session starts from evidence.

Reports and AI feedback help you decide what to practise next. They do not guarantee rank improvement, admission or selection. Rank and college predictors are not offered.