AlShaheen الشاهين

An AI coach for self-practice on the Qudrat (General Aptitude Test)

“I practise a lot, but I keep repeating my mistakes and don’t know what to review.”

Students study from booklets, channels and videos, and solve a lot of questions — but they don’t know why they got one wrong or what to review today, so the same mistake follows them to exam day.

AlShaheen gives the mistake a name and brings the same pattern back with new questions, until the student solves it correctly days later without help.

Try a question — no sign-up

The live demo is in Arabic, like the test itself.

Who it’s for

How it works

  1. Solve a question in the style of the test, then say how sure you were: confident, unsure, or guessed.
  2. Get it wrong and you see your mistake by name, in a student’s words — not just “wrong”.
  3. Two hints before the solution, then a colour-coded explanation, step by step.
  4. A similar question with a different context and wording — not just different numbers.
  5. Two days later the pattern returns. Solve it with no hints and the mistake moves to “mastered”.

240 km ÷ 3 hours = 80 km/h

Each element keeps the same colour from the first step to the last.

What you get

Today’s session, ready

8 quantitative and verbal questions picked from your mistakes and due reviews. Short on time? “I only have 5 minutes”.

Your top 3 mistakes

Named, with how often they repeat — not “weak in percentages”. A “mastered” counter that never goes down.

Diagnostic or from scratch

10 questions in 7 minutes to order your start, or begin directly from the first skill.

Foundations

When a pattern needs pre-aptitude basics, a short lesson comes first, clearly labelled “general foundations”.

Return without guilt

Away for a few days? No “you missed”. You restart from your most important mistake with a shorter session.

Report a problem

A button on every question, with a reply within 24 hours.

What it covers

Every question is newly generated in the style of the test; no circulated questions are used. The demo is live now; the full first version is in development.

Powered by Claude

Claude by Anthropic

Questions and explanations are written in advance by Claude, Anthropic’s AI model. No question reaches a student before it is checked:

  • Quantitative: every answer is computed by code, not by the model. A second, independent model solves the question without seeing the answer; any disagreement discards the question.
  • Verbal: structural checks in code, an independent judge model, then a top-scoring student reviewer reads every question.

AI for language, code for truth. AlShaheen is an independent project; it is not an Anthropic product and is not affiliated with Anthropic.