AI quiz generators by source

QuizGen generates quizzes from PDFs, images, YouTube captions, webpages, video files, keywords, and text.

Compare input methods, supported formats, and limitations by source, with real generation cases and sample questions.

Quiz creation and editing require an account and sign-in.

View generation cases

Quiz creation process

  1. 1. Choose a source type

    Select the generator that matches the format and purpose of your material.

  2. 2. Extract key content

    Organize text, captions, or visual information into content that can support questions.

  3. 3. Review and share

    Edit the generated questions, publish a quiz link, and review submissions.

Quiz generation by source

PDF, image, and YouTube cards link to detailed guides. URL, video, and keyword cards link to input and question examples below.

PDF Quiz Generator

Document-based quiz generation

  • PDF upload
  • Up to 25MB
  • Extract text before generation
View detailed guide

Image Quiz Generator

Image-based quiz generation

  • PNG, JPEG, GIF, WebP
  • Up to 10 images
  • Up to 5MB per image
View detailed guide

YouTube Quiz Generator

Caption-based quiz generation

  • YouTube URL input
  • Caption language selection
  • Caption-based questions
  • Available captions are required.
View detailed guide

URL Quiz Generator

Enter the URL of an article, blog post, or learning webpage to create draft questions from its text.

  • Webpage URL input
  • Text extraction
  • Preview before generation
  • Pages requiring login or restricting access may not be imported.
Input and question example

Video Quiz Generator

Audio transcription-based quiz generation

  • MP4, MOV, WEBM, and more
  • Up to 500MB
  • Transcribe audio before generation
Input and question example

Keyword Quiz Generator

Create draft questions from a topic keyword without a source document. Use optional sub-keyword suggestions to refine the topic.

  • Direct keyword input
  • Sub-keyword suggestions
  • Keyword up to 50 characters
Input and question example

Real generation cases

Review results generated from actual PDF, image, YouTube, webpage, and video file sources, along with source-check notes.

QuizGen case

AI lecture review

Five questions generated from a Whisper-1 lecture transcript, with one replaced after review.

Source material
An actual lecture video file uploaded directly
Generated result
Five final review questions checked against the transcript
Question format
3 multiple-choice and 2 short-answer questions (manually reviewed)
View case
QuizGen case

Anti-smishing card news

Text from seven card-news images was extracted, checked against the source, and curated into a mixed-format quiz.

Source material
Seven Korea Consumer Agency card-news images
Generated result
Five questions about smishing methods and prevention
Question format
3 multiple-choice and 2 true/false (O/X) questions (curated review)
View case
QuizGen case

VGGNet paper

A research paper turned into key-facts questions about VGG architecture and evaluation.

Source material
VGGNet research paper PDF
Generated result
Key-fact questions about VGG architecture and ILSVRC metrics
Question format
Multiple-choice questions with review notes
View case
QuizGen case

Neural-network video

Full captions summarized into a neural-network fundamentals quiz.

Source material
Full captions from a neural-network lecture
Generated result
Five-question neural-network fundamentals quiz
Question format
Multiple-choice questions with LaTeX math
View case
QuizGen case

Wikipedia linguistics page

Basic linguistics concepts from a public webpage turned into five O/X (True/False) questions.

Source material
Wikipedia linguistics article URL
Generated result
Five O/X (True/False) questions about basic linguistics concepts
Question format
True/false (O/X) questions checked against the source
View case

Source inputs and sample questions

These static examples show the kind of question each source type can produce.

PDF

Source paper PDF

Very Deep Convolutional Networks for Large-Scale Image Recognition. The 14-page paper was used to generate questions about VGG architecture, ILSVRC metrics, and feature extraction.

What is a central VGG design idea?

Sample answer: Using small 3×3 convolution filters while increasing depth

Image

Sample image content

A photosynthesis diagram shows light, carbon dioxide, and water entering a leaf, while oxygen and glucose are produced.

What substances are produced during photosynthesis?

Sample answer: Glucose and oxygen

YouTube

Source YouTube material

The full captions from “How to train Neural Network (신경망 학습 원리, 역전파)” were summarized into questions about weights, loss, gradient descent, momentum, and automatic differentiation.

What is the calculation order in a learnable layer?

Sample answer: Multiply by W, add b, then apply ReLU

URL

Sample webpage excerpt

Climate change refers to long-term shifts in temperature and precipitation patterns. A major cause is increased greenhouse gases, and effects include sea level rise and extreme weather.

What is one major cause of climate change?

Sample answer: Increased greenhouse gases

Video file

Sample transcript excerpt

During photosynthesis, plants use light energy to convert carbon dioxide and water into glucose and oxygen. This process mainly happens in chloroplasts in the leaves.

Where does photosynthesis mainly happen in plant cells?

Sample answer: Chloroplasts

Keyword

Sample keyword

Photosynthesis: a biology concept related to plants using light energy to make nutrients, chloroplasts, carbon dioxide, water, glucose, and oxygen.

Which materials are needed for photosynthesis?

Sample answer: Carbon dioxide and water

Frequently asked questions

You can create quizzes from seven source types: keywords, text, PDFs, webpage URLs, YouTube videos, video files, and images.

PDFs use the document text, URLs use the webpage body text, YouTube uses captions, video files use a speech transcript, and images use visible text and visual content. Keywords generate questions from a topic and suggested sub-keywords without source text.

Six types: multiple choice, O/X (True/False), short answer, essay, fill in the blank, and matching.

Up to 10 questions per quiz and up to 10 questions per question type. Matching questions cannot be combined with other question types.

Yes. Quiz generation language and keyword suggestion language can each be set to Korean or English.

From the Plus plan, you can write requirements in the additional instructions of the quiz settings and pass them to the AI. Difficulty, generation language, question types, and question counts can be chosen on the Free plan as well.

Quiz creation and editing require logging in to an account. Without logging in, you can try the question-taking, scoring, answer, and explanation flow in the AI quiz demo.

Quiz demo without login

Answer sample questions and check your score, answers, and explanations.

Try the AI quiz demo