MYP teachers build most of their assessment materials from scratch. There is no shared question bank. The IB framework is precise about which command terms belong at which level band, and Criterion B in Biology is not the same as Criterion B in Mathematics. Getting those details wrong means setting the wrong cognitive bar for students, and most teachers spend 45 to 90 minutes building a single criterion paper before they've written a marking note.
PrepAiro's Assessment Generator solves the specific problem, not the generic one. It enforces command terms by level band, loads the correct strand descriptors per subject and criterion, and exports separate student and teacher PDFs, all in two structured steps.
💡 Key Takeaways
- MYP teachers spend 45-90 minutes building a single criterion assessment from scratch, with no shared question bank available.
- Command term enforcement is built into the Level Target selector: Level 1-2 locks in State/Identify/Define; Level 7-8 locks in Evaluate/Justify/Discuss.
- Criterion strand descriptors load per subject: Biology Criterion B and Mathematics Criterion B show different strands in the sidebar.
- Each output card can be kept, edited inline, or regenerated individually without affecting other cards.
- The tool exports separate student and teacher PDFs and connects directly to Rubric Maker with no re-entry.
Why Does Building One MYP Assessment Take So Long?
The IB MYP framework requires teachers to assess against subject-specific criteria, each with distinct strand structures. MYP subjects use four criteria (A, B, C, D), but the strand descriptors inside each criterion differ by subject. Biology Criterion B covers "Inquiring and designing" with three specific strands. Mathematics Criterion B covers "Investigating patterns" with a completely different set of three strands. A teacher cannot reuse question structures across subjects without re-checking alignment.
Command terms add another layer. The IB specifies different cognitive demands by achievement level: levels 1-2 require terms like State, Identify, and Define; levels 5-6 require Analyse and Suggest; levels 7-8 require Evaluate, Justify, and Discuss. Teachers who write a "Discuss" question targeting a level 3-4 band have set a misaligned cognitive bar, and standard rubric language will not catch it at the question level.
The only purpose-built alternative is AssessPrep, which holds a strong question bank. It is locked inside ManageBac school subscriptions, though, which means individual teachers or schools not on ManageBac cannot access it. ChatGPT produces coherent question text but has little depth.
Step 1: Content Setup
Step 1 collects the subject context. The Year Group dropdown covers MYP 1 through 5. The Subject dropdown includes Biology, Chemistry, Mathematics, Physics, and Standard Mathematics. The Topic field is free text, so you can be as specific as your unit requires.
The file upload accepts PDF or PPTX files up to 20MB. This is optional: questions generate from subject and topic alone if no file is attached. Teachers who have built slide decks or reading packets for a unit can upload them and get questions grounded in their own content rather than a general curriculum model.
If you have set a teaching profile in PrepAiro, the subject and year group fields pre-fill automatically.
Step 2: Question Parameters
Criteria checkboxes with a live strand sidebar. Criteria A, B, C, and D appear as checkboxes. When you select a criterion, the sidebar updates to show that subject's strand descriptors. Select Criterion B for Biology and you see "Inquiring and designing" with its three strands. Switch to Mathematics and the same Criterion B shows "Investigating patterns" with a different strand structure.
Level Target: enforcement, not labelling. The Level Target selector offers five options: 1-2, 3-4, 5-6, 7-8, and Mixed. This is not a tag applied after generation. The tool uses the selected level band to determine which command terms are valid. Level 1-2 produces questions using State, Identify, and Define. Level 7-8 produces questions using Evaluate, Justify, and Discuss. You cannot get a "Discuss" question from a 1-2 band selection.
Question Types. MCQ, Short Answer, Extended Response, and Data Processing. Data Processing questions embed stimulus material automatically: graphs, tables, or diagrams are included in the card. You do not need to source or attach them separately.
The Output: Question Cards
Each generated question appears as a card showing the question stem, criterion tag, command term, achievement level, mark count, and marking notes. Everything needed to review alignment is visible at the card level.
Three actions sit on each card. Keep adds it to your paper with a green border. Edit opens an inline editor to change the stem, mark count, or marking notes. New regenerates only that card using the same parameters, leaving all other cards untouched. A running counter shows your current state: "5 kept · 1 edited · 2 replaced."
Test Paper mode unlocks once you have kept three or more questions. You add a title, instructions, and a time limit. Two exports are available: the Student PDF (question stems and answer space, no marks or tags) and the Teacher PDF (questions, answers, marking notes, and criterion tags). Both generate from the same kept set.
Connected to Rubric Maker
A "Generate Rubric" button appears in the output panel once questions are kept. Clicking it carries the subject, topic, and selected criteria into Rubric Maker with no re-entry.
How Does It Compare to the Alternatives?
AssessPrep holds the strongest existing question bank for IB. It is built for assessment delivery inside ManageBac. Schools without a ManageBac subscription cannot access it, and it does not generate new questions via AI.
ChatGPT produces fluent question text with no MYP strand descriptors, no command term enforcement by level band, and no past paper sourcing. A question that looks correct can still use the wrong command term for the intended level, and ChatGPT will not flag it.
MagicSchool is a broad AI teaching assistant with no IB criterion awareness.
Knowt is student-facing and is not a criterion-based assessment creation platform.
The structural difference is command term enforcement and strand-level criterion awareness. Those two features are what make MYP assessment creation technically difficult, and they are what generic AI tools skip.
Frequently Asked Questions
Can I use my own uploaded materials as the question source?
Yes. Upload a PDF or PPTX file up to 20MB in Step 1. When a file is present, the tool generates questions from your materials. The same criterion, command term, and level band rules apply.
What happens if I select Mixed for both Level Target and Difficulty?
The tool generates questions across the full range of command terms and difficulty levels. Each card still shows its individual level target and command term, so you can see the spread and remove cards that don't fit your intent.
Does the tool work for Standard Mathematics separately from Mathematics?
Yes. Standard Mathematics appears as its own subject. Criterion strand descriptors and question generation treat it as distinct from the general Mathematics pathway.
Can I edit a question after I've kept it?
Yes. Kept questions remain editable. Click Edit on any card, including those already marked as kept, to open the inline editor. The running counter updates to reflect the edit.
Get Started
Open a new assessment, set your subject and year group in Step 1, select your criteria and level target in Step 2, and generate your first set of cards in under two minutes. Teachers who have set a teaching profile will find subject and year pre-filled on every session.
