Grade AI

AI test questions from your own notes: how it works and what to check

How Grade AI drafts multiple-choice questions from a topic, photos of notes, PDFs or Word documents, what the AI receives and what it does not, and how to check the questions before printing.

· 4 min read · Česky

In short: In Grade AI you can let AI draft the multiple-choice questions for a test, either from a topic you type or from your own materials: photos of notes, PDFs, Word documents or text files. You can say who the test is for and what to focus on. The questions land in the editor, where you check and change them before anything is printed. Using the AI is optional.

What you can give the AI

Questions made from your own materials stay close to what you actually taught, in the words you used. A topic alone gives more general questions.

What you can set

Check before you print

AI drafts are a starting point, not a finished test. Before printing, go through every question:

  1. Is the right answer right? AI can be confidently wrong, especially with numbers, dates and names.
  2. Is exactly one answer right? A second answer that is "also partly true" makes a question unfair.
  3. Are the wrong answers plausible? Good distractors are mistakes students really make. Obviously silly options make a question too easy.
  4. Does it match what you taught? Remove questions about things the class has not covered.
  5. Is the wording clear for this age? Shorten long questions; the sheet has limited space per question.

Everything can be edited in the app: the question, each of the four answers and which one is correct.

What the AI receives, and what it does not

Grade AI uses Anthropic's AI to write the questions. It receives only what you give it for that purpose: the topic, your note about focus and audience, and the materials you upload. It does not receive your name, your email, or anything about students: no names, no answers, no results and no photos of answer sheets.

Uploaded files are not stored by Grade AI. They pass through to the AI inside the request and are gone once the questions are generated. Under its terms, Anthropic does not use data sent through its API to train models and deletes it within 30 days. The full details are in the privacy policy.

Limits

Frequently asked questions

Do I have to use the AI?

No. You can write every question yourself. The AI is a shortcut, not a requirement.

Can the AI make mistakes?

Yes. That is why the questions go to the editor first. Check the correct answer of every question before you print.

Are my notes stored somewhere?

Not by Grade AI. The files are sent to the AI inside the request and are not saved in the app. Anthropic deletes API inputs within 30 days and does not train on them.

Can it make questions in my language?

Yes. The AI writes in the language set for AI questions, which is the app's language unless you change it. It can be any of the app's ten languages, so a language teacher can keep the app in English and get the questions in Spanish.

Try it on your next test.

Set up a class, print the sheets, see how long it takes. Scanning happens on the phone. Building and printing work in the browser too.

Get the app

More articles