Anyone who has sat a badly written multiple-choice exam knows the trick. You do not need to know the answer. You need to notice that most of the options are absurd, one is suspiciously longer than the rest, and the answer is almost never "none of the above."
That is a test of test-taking, not of knowledge. And it is the specific failure that separates a useful generated quiz from a decorative one.
The wrong answers are the whole product
A multiple-choice question is only as good as its distractors — the four incorrect options. Get them right and the question measures whether someone understood the material. Get them wrong and it measures whether someone can spot a joke option.
Good distractors have a particular quality: each one represents a mistake a real person might actually make after watching the video. A near-miss on a definition. The right concept attached to the wrong example. A figure from a different part of the talk. Someone who understood the material rejects them all quickly; someone who half-understood it hesitates. That hesitation is the measurement.
XYZ Youtube generates the four incorrect options from the video's own content for exactly this reason. Terminology the speaker used, adjacent claims they made, numbers they mentioned elsewhere. The distractors are wrong, but they are wrong in ways that come from the material rather than from thin air.
What you get
Paste any YouTube URL — or type xyz in front of one — click Analyse, choose Quiz, and pick a difficulty level, and you get questions with five options each, one marked correct, and a short explanation. Length scales with the video and the level: a fifteen-minute explainer has less distinct testable material than a ninety-minute lecture, and the output reflects that honestly rather than padding to hit a number.
The output is plain text, which means it goes wherever you need it — a document, a quiz platform, a learning management system, a slide.
Vetting a generated set in about a minute
Worth doing before anything graded. Three checks catch nearly everything:
Read the distractors, not the answers. If any question has an option that is obviously silly, that question is now four-option and easier than intended. Rewrite or drop it.
Check every number against the video. This is where auto-captions fail most often. "Fourteen point two percent" and "forty point two percent" are one transcription error apart, and a question built on the wrong one is confidently, invisibly wrong.
Watch for questions with two defensible answers. Occasionally a speaker says something loosely enough that a distractor is also arguably true. Easiest fix is deletion.
A minute of this is the difference between a quiz you can hand to thirty students and one you cannot.
Where MCQs are the right instrument
Trivia and games. Documentary and history videos turn into pub-quiz rounds with almost no effort. The format does not care whether the content is academic.
Onboarding and compliance. If your team is required to watch a training video, a short MCQ set is the standard evidence that they did. Generating it from the video itself keeps the questions aligned to what was actually taught.
Self-testing at speed. Multiple choice is fast to take, which makes it good for repeated low-effort checks over several days — the pattern that actually moves material into long-term memory.
Classroom assessment. Teachers building quizzes from assigned videos have the deepest version of this workflow, including how to handle a class that has not watched the homework.
When to use something else
MCQs test recognition. They are poor at testing whether you could explain an idea unprompted, and useless for testing whether you could apply it somewhere new.
If you want open questions you have to answer in your own words — better for flashcards, spaced repetition, and interview prep — use Q&A pairs instead. For the broader case for testing yourself on video content at all, and how to fit it into a study routine, start with generating a quiz from any video.
Pick a documentary you watched recently and generate a set. Reading the three wrong answers tells you more about how the format works than any explanation. Try it.