Pattern analysis for university exams
Your next exam paper, predicted.
Upload past papers and get back a mock paper with an honest probability on every question. Frequency math decides the numbers — AI only writes the words.
Why it works
Built on recurrence, not vibes
Numbers you can defend
Every probability is frequency math over real past papers — appeared 4 of 5 years means 80%. The AI never invents a percentage.
Reworded ≠ new
Examiners rephrase the same question for years. Semantic matching merges the rewordings so the counts stay honest.
Your exam's exact format
The blueprint engine learns which slots repeat — 1(a) shorts, 3(b) numericals — and detects concepts that rotate year over year.
Cold-topic warnings
Upload the syllabus and heavily-weighted topics that were never asked get flagged before they ambush you.
Ranked, not random
A frequency-ranked list of what to study first, with the years each question appeared and its typical marks.
Predict without quota
The paper can be assembled from historical phrasings alone — no AI call, no API key, same probabilities.
Workflow
Three steps to a predicted paper
Upload past papers
PDFs or photos. Text is extracted and every question — including 2(a), 2(b) sub-parts — is parsed into a structured bank.
Analyze the pattern
Reworded repeats are merged by meaning, then each question’s recurrence over the years becomes its probability.
Get the predicted paper
A mock paper in your exam’s exact format, a stamp on every question, and surprise flags for untested syllabus topics.
Stop guessing what to study tonight.
Three or more years of past papers is all it takes.
Open dashboard