Here is an uncomfortable finding that most students discover the week before an exam rather than the week after the lecture: watching is not learning. Rewatching in particular produces a strong feeling of mastery and very little of the real thing. The material is familiar, familiarity feels like knowledge, and the gap only shows up when someone asks you a question.
Testing yourself is the fix. It has been the fix in the cognitive science literature for decades — retrieval practice beats review by a wide margin. The problem was never that people did not know this. The problem is that writing good questions about a lecture you have just watched takes longer than watching it again.
The workflow, end to end
- Take the YouTube URL and type
xyzin front of it:xyzyoutube.com/watch?v=.... Or paste it at xyzyoutube.com. - Click Analyse.
- Choose Quiz, then pick a difficulty level.
You get multiple-choice questions, each with five options, one correct answer and a short explanation, drawn from what was actually said in the video. Longer videos and harder levels produce more questions. No account needed for your first one.
That is the whole thing. The interesting part is what you do with it.
Pick the level that matches what you want to find out
The three levels are not the same questions made longer or shorter. They test different things.
Beginner asks about facts, definitions, names and figures that were stated outright. Every question stands on a single point, and the wrong answers are clearly wrong once you know the fact. Use it to check that you were paying attention.
Intermediate asks about concepts, cause and effect, the reason behind a recommendation, and applying a stated rule to a new case. Some questions need two parts of the video connected. This is the default, and it is the right one for most revision.
Expert asks about implications, trade-offs, limits and edge cases. Questions are built to be answered by judgement rather than memory, and the wrong answers are near-misses — true in general, wrong under the condition being asked about. Difficulty comes from depth, never from obscure trivia: everything is still answerable from the video alone.
A useful pattern is to run Beginner immediately after watching, and Expert a week later. The first tells you whether you absorbed the material; the second tells you whether you understood it.
Use it before you review, not after
The instinct is to watch, revise, then quiz yourself to confirm you have it. Flip that. Generate the quiz first and take it cold, immediately after watching, before any revision.
You will do worse. That is the point — the questions you miss are a map of what did not stick, and you now know exactly where to spend your revision time instead of rereading everything at uniform effort. A cold score of 4/10 that directs you to the right three topics is worth more than a warm 9/10 that tells you nothing.
Where a generated quiz beats a written one
Lectures you attended but did not absorb. Recorded classes go up after the session. A quiz on the recording is a five-minute check on whether the hour was worth anything.
Certification and interview prep. A lot of prep material lives in long conference talks and technical deep-dives. Turning each into a quiz converts passive viewing into something closer to practice questions. Pair it with Q&A pairs for the open-ended half.
Teaching. If you assign a video for homework, a quiz is how you find out who watched it — and, more usefully, which part confused everyone. See how teachers use this for the classroom workflow.
Reading groups and team training. Anywhere a group is meant to have watched the same thing, a shared quiz surfaces the parts that need discussion.
What the questions can and cannot test
Multiple choice tests recognition and recall of stated content. It does that well and cheaply. The questions come from the transcript, so anything the speaker said clearly is fair game.
What it cannot test is synthesis — whether you could apply the idea to a new situation, argue against it, or connect it to something from a different lecture. No multiple-choice question does that, generated or hand-written. For that, use Q&A pairs, which pose open questions you have to answer in your own words, or work from a mind map that shows how the pieces relate.
A practical note on trust: if you plan to use a generated quiz for anything that carries a grade, read it first. Auto-captions occasionally mangle a number or drop a negation, and a question built on a mangled line will be confidently wrong. Thirty seconds of review is enough to catch it. For a deeper look at how the questions are structured — and why the wrong answers matter as much as the right one — see multiple choice questions from video.
Try it on something you just watched
The best test of whether this is useful is uncomfortable and takes two minutes: pick a video you watched in the last week and felt good about. Generate the quiz. Take it honestly.
Most people are surprised. Find out.