Generate Personalized Exam Study Roadmaps

ExamPrep AI converts a syllabus, past exam papers, or study notes into a prioritized study roadmap, and can also generate flashcards, a predicted exam paper, or an exam-readiness score. It uses a selective-reading rule internally - routing to only the section matching the student's request (Full Roadmap, Theory, Numerical, MCQ, Coding, Lab, Flashcards, Predicted Exam Paper, or Readiness Dashboard) plus a small always-loaded Shared Foundations block, rather than loading every mode for a focused request.

Shared Foundations define a universal difficulty scale - Easy (define/state/list/identify, guaranteed marks, study first), Medium (explain/describe/compare/calculate, mid-paper marks), Hard (derive/prove/optimize/analyze/design, score separators, study last) - always presented Easy to Medium to Hard, never reversed. Intake asks once for at least a syllabus, past papers, notes, or subject name plus university; confirms the detected course code if OCR confidence is under 80%; and asks how much time is available, defaulting to Standard Mode (6-12 hrs) if unanswered. Four study modes scale the workload: Emergency (1-2 hrs, Easy only, top 10 questions), Sprint (3-5 hrs, Easy+Medium, top 25), Standard (6-12 hrs, all difficulties, full roadmap), and Advance (days+, daily schedule plus mock papers). A syllabus guardrail maps every question to a syllabus unit at 70%+ match to tag it [IN SYLLABUS], never generates content for topics absent from the uploaded syllabus, and flags out-of-syllabus items for the student to approve before inclusion. A probability score ranks question importance as (Frequency x 0.40) + (Recency x 0.30) + (Unit Weight x 0.20) + (Marks x 0.10), where frequency is appearances over max appearances, recency scores higher for the last 2 years, unit weight scores core topics above electives, and marks weight scores higher-value questions higher.

Full Roadmap Mode extracts all questions with year/source, confirms the count and type breakdown with the student, classifies each into one of five types (Theory, Numerical, MCQ/True-False, Coding, Lab) by keyword signal, builds ranked tables per type with times-seen/marks/difficulty/unit/priority, generates notes ordered Easy-across-all-types before Medium before Hard, produces a per-unit coverage tracker distinguishing past-paper-sourced from AI-predicted content (labeled [PREDICTED - not from past papers]), and then offers flashcards, a predicted exam paper, or a readiness dashboard. Each of the five question types has its own three-tier note template (Easy/Medium/Hard) tailored to that type - for example Numerical notes include a worked example with common-mistake and memory-hook fields, Coding notes include a trace table for hard-tier trace/optimize/debug tasks, and Lab notes include a viva Q&A block for hard tier. Flashcards render one card per question with type/difficulty emoji, question, short answer, and a key formula/term. The Predicted Exam Paper assembles a labeled mock paper (explicitly marked "AI PREDICTION - Not official") with short/objective, medium-answer, and long-answer sections matching the difficulty tiers. The Exam Readiness Dashboard reports per-type, per-difficulty percentage coverage, an overall preparedness percentage, an estimated marks range, strong vs. weak topic lists, and a confidence rating tied to how many past papers were supplied.

Before every output it runs quality checks: every note maps to a syllabus unit, difficulty order is never reversed, worked numerical examples compute correctly, code snippets are syntactically valid, notes are readable in 2-5 minutes, no fact appears that isn't in the uploaded materials, and the OCR-detected course code has been confirmed by the student. It also has scripted error responses for missing syllabus, single-past-paper low confidence, OCR failure, out-of-syllabus questions, mixed subjects, no time given, and missing numerical/coding questions.

Use this skill to convert a syllabus, past papers, or study notes into a prioritized roadmap, to focus on a specific question type (Theory, Numerical, MCQ, Coding, Lab), to create flashcards, a predicted exam paper, or a readiness check, or for last-minute revision or a deep-dive into important topics. This skill supports study planning and revision but cannot guarantee actual exam questions, marks, grading outcomes, or instructor expectations; probability scores are heuristics that degrade with sparse, outdated, or incomplete inputs; it should not fabricate syllabus coverage and must ask before adding predicted content when source material is missing or ambiguous; it is not a substitute for official course guidance, accessibility accommodations, academic-integrity policies, or instructor feedback; and it should not request or process private student records beyond the study material needed for the current task.

Students preparing for an exam who have a syllabus and/or past papers and want a prioritized, time-boxed study plan with type-specific notes, flashcards, a mock paper, or a readiness estimate rather than generic study advice.