Do AI Quiz Makers Work on Scanned PDFs?

The short answer: it depends entirely on how the tool reads your file, and the two kinds of tool behave completely differently. A quiz maker built on a vision capable model looks at the page the way you do and writes questions straight from a scan, a photo or a screenshot, with no preparation step. A quiz maker built only on text extraction opens the same file, finds no text layer, and returns nothing usable. Same PDF, opposite outcome. Before you blame the file, find out which kind of tool you are using.

This trips people up because a scanned PDF and an exported PDF look identical on screen. Open a chapter you photocopied and a chapter exported from Word and they are both just pages. Underneath they could not be more different: the exported file stores the words as characters, and the scan stores a picture of the words. Whether that difference matters to you is a fact about your quiz tool, not about your document.

How to tell whether your PDF is a scan

Open it and try to select a sentence with your cursor, or search for a word you can plainly see on the page. If the text highlights or the search finds it, the file has a text layer. If nothing selects and the search comes back empty, you are looking at images.

Anything that arrived through a scanner, a photocopier, a phone camera or a screenshot is an image. So are most older textbooks, archived documents and the handouts that have been photocopied so many times the edges have gone gray.

The two kinds of quiz maker

PDFQuiz is in the first row. Scans, photos and image files are sent to the model as images and read visually, so making a quiz from a scanned PDF is the same three clicks as any other upload. There is no OCR step to run, nothing to retype, and no separate searchable copy to keep track of.

If you have used a quiz tool that came back with vague questions from a scanned chapter, that tool was almost certainly in the second row. The symptom is distinctive: the questions are about the general subject rather than about anything specific on your pages, because the tool had nothing from your document to work with and fell back on what it already knew about the topic.

What actually determines quality on a scan

With a vision model the constraint moves from file format to legibility. If you can comfortably read the page on your own screen, the model can too. These are the things that genuinely change the result.

• Focus and resolution. A sharp phone photo beats a low resolution scan. Blurry small print is the most common cause of weak questions.
• Angle. Photograph the page flat and square on. A page shot at an angle, or a thick book that curves away at the spine, loses text into the shadow.
• Lighting. Even light, no glare, no hand shadow across the middle of the page.
• Contrast. A tenth generation photocopy where the letters have started to fill in is hard for anyone to read, and the questions will reflect that.
• Completeness. Capture the whole page including the diagram captions. Cropped edges quietly remove content from the quiz.

Notice that every item on that list is about how well the page can be read, not about what kind of file it is. That is the practical difference vision makes.

Is there any reason to run OCR first?

Sometimes, though not for the reason people usually think. You do not need OCR to generate questions. You might still want it when the scan is genuinely poor and you would rather see and correct the recovered text before anything is built on it, or when you need the document searchable and copy-pasteable for its own sake, or when you are processing a large archive and want one clean text version to work from afterwards. For those jobs a dedicated engine like DocuOCR document data extraction gives you output you can inspect. For simply turning a scanned chapter into a quiz, it is a step you can skip.

Do AI quiz makers work on handwritten notes?

Yes, when the handwriting is legible, and the same rule applies: neat print style writing on lined paper reads well, while cramped cursive or faint pencil does not. Photograph the pages flat in good light and upload them like any other file. Where handwriting gets read imperfectly you will usually see it in the questions, and a question built on a misread word is obvious enough to delete in a second. Turning a semester of handwritten pages into practice questions from your notes works better than most people expect.

Can I quiz from a photo of a textbook page?

Yes. A phone photo of a page is a scan captured by your camera, and it uploads exactly the same way. Lay the book flat so the spine does not curve the text away, fill the frame with the page, avoid glare from an overhead light, and shoot two facing pages separately rather than as one wide image.

How many questions can a scanned chapter produce?

Roughly three to five solid questions per page of substantive content, which puts a ten page scanned chapter somewhere around thirty to fifty. That is the same yield as any other source. Legibility affects quality rather than quantity: a clean scan and a native PDF of the same chapter produce comparable questions, while a dark, skewed photo produces fewer usable ones.

What if a scan still produces bad questions?

Re-capture the worst pages before you do anything else, because that fixes it most of the time. Beyond that, check that you uploaded every page rather than a partial selection, and be aware that pages which are mostly diagram carry very little testable text no matter how well they are photographed. If a chapter leans heavily on figures, pair it with the lecture notes covering the same material so the model has something to write from.

The practical summary

Ask what your quiz maker does with an image. If it reads pages visually, a scan is no different from any other upload and your only job is a legible capture. If it only extracts text, you will need to OCR every scanned file before it is useful. Photocopied handouts, older textbooks, printed notes and archived documents are some of the most common material people want to test themselves on, and none of it should require a preparation step. Upload the scan to the PDF to quiz tool and see what comes back, or start from a study guide built from the same pages if you want the concepts and glossary before the questions.