How to Make a Quiz From a Scanned PDF

The short answer: upload the scanned PDF, set how many questions you want, and generate. The pages are read visually, the same way you read them, so there is no OCR step to run first and nothing to retype. A ten page scanned chapter takes about a minute and comes back as questions with the answer key marked. The one thing that matters is legibility: if you can read the page comfortably on your own screen, so can the model.

This surprises people who tried it a few years ago, when quiz tools worked purely by pulling text out of a file and a scan came back empty. That constraint is gone on any tool built on a vision capable model. It is worth knowing the difference, because plenty of guides still open by telling you to OCR everything first.

How to make a quiz from a scanned PDF, step by step

1. Gather every page you want covered. If the material is split across a photocopied handout and your own notes, put them in one upload rather than doing two runs, so the questions span the whole topic.
2. Check the pages are legible. Open the file and read a paragraph at normal zoom. If you are squinting, re-scan or re-photograph those pages now. This is the only preparation worth doing.
3. Upload the file. Scanned PDFs, phone photos, screenshots and image files all upload the same way as an exported PDF.
4. Set the number of questions and the format. Multiple choice, true or false, short answer or a mix, plus how hard the questions should be.
5. Generate, then read the draft. Anything built on a word that got misread will look odd immediately, and deleting it takes a second.
6. Export. Take the quiz and the answer key as a PDF to print, or as an editable file to paste into your LMS.

How do I know if my PDF is a scan?

Try to select a sentence with your cursor, or search the document for a word you can plainly see on the page. If nothing highlights and the search finds nothing, each page is stored as an image rather than as text. That is a scan. Anything from a scanner, a photocopier, a camera or a screenshot falls into this category, along with most older textbooks and archived documents.

It is useful to know which you have, mostly so you can predict how other software will behave with the file. For generating questions here, it changes nothing about what you do.

What actually changes the quality of the questions

File type is not the variable. Capture quality is. These are the things worth fixing before you upload.

Every row is about how readable the page is. None of them is about the file format, which is the practical point.

Do I need to OCR a scanned PDF before making a quiz?

No. The pages are read as images, so a scan goes straight in. OCR is still worth running for other reasons: when a scan is poor and you want to see and correct the recovered text yourself before anything is built on it, when you need the document searchable and copy-pasteable, or when you are digitizing an archive and want one clean text version to keep. A dedicated engine like DocuOCR document data extraction handles that properly and gives you output you can inspect. For the specific job of turning a scanned chapter into questions, it is an extra step that buys you nothing.

Does it work on handwritten notes?

Yes, when the handwriting is legible, and neatness matters more than anything else. Clear print style writing on lined paper reads well. Cramped cursive, faint pencil and heavily abbreviated notes read less well, and you will see that in the questions. Photograph the pages flat under even light and upload them like any other file. A semester of handwritten pages turns into practice questions from your own notes more reliably than most people expect.

Can I make a quiz from a photo of a textbook page?

Yes. A phone photo is a scan captured by your camera and uploads identically. Lay the book flat so the spine does not curve the text away, fill the frame with the page, avoid overhead glare, and shoot facing pages separately rather than as one wide image. Ten pages photographed properly beats thirty photographed in a hurry.

How many questions will a scanned chapter give me?

Around three to five solid questions per page of substantive content, so a ten page chapter usually lands between thirty and fifty. That matches what the same chapter would yield as a native PDF. A poor capture reduces the count rather than quietly producing wrong questions, because pages that cannot be read simply contribute less.

Why do some quiz tools fail on the same file?

Because they read text rather than pages. A tool that only extracts a text layer opens a scan, finds nothing, and either errors out or falls back on general knowledge about the subject. That fallback is the giveaway: the questions sound plausible but do not reference anything specific to your document. If you have seen that behavior, it was the tool, not your scan. There is more detail in whether AI quiz makers work on scanned PDFs.

Getting more out of scanned material

Photocopied handouts, older textbooks, printed notes and archived documents are some of the most common material people need to test themselves on, and it is usually the material with no digital version anywhere. Once a scan is in, you are not limited to a quiz: the same upload can produce a study guide with the concepts, glossary and review questions for the same pages, or a set of flashcards. Start with the scanned PDF to quiz tool and see what your worst photocopy comes back with, which is usually more than you would guess.