Claude Code turns a CS50 transcript into an HTML quiz
@RoupenMD used Claude Code to turn a CS50 transcript into an HTML quiz that strikes through covered transcript passages. The result, reported by the poster, was 84 practice questions.
x.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Turn [course transcript] into an HTML practice quiz. Link each question to its source passage and strike that passage through after I answer. Keep the transcript visible so I can see which material I have covered. Flag questions whose answers are unclear in the supplied text.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source poster / publisher
- @RoupenMD
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- 84 practice questions in an HTML quiz with transcript coverage struck through as questions are answered, reported by the poster. Question accuracy and learning gains were not independently tested.
- Human involvement
- The poster selected and supplied the transcript and specified the interaction. The post does not describe a question-by-question accuracy review.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- The learner copied the available CS50 transcript into Claude Code.
- They requested an HTML quiz linked to the transcript passages covered by each answer.
- Claude Code generated the quiz; the poster reported 84 practice questions and shared a screenshot.
@RoupenMD had finished watching Harvard's CS50 course on edX and wanted a way to practise the material rather than simply watch it. Harvard made the lecture transcript available, so the poster copied it into Claude Code and requested an HTML file containing questions to test their understanding.
The requested interaction connected each answer to the source material. When the learner answered a question, the corresponding part of the transcript would be struck through. This made the transcript a record of material covered during practice, rather than a separate document to consult manually.
The result was 84 practice questions, reported by the poster, who also shared a screenshot of a question. The post describes the output arriving a few minutes later. It does not provide a downloadable HTML file, a correctness evaluation of the questions or evidence of improved exam performance.
The human supplied the course material and the behaviour to build, then used the resulting questions for study. Claude Code's role was generating the interactive study artifact, not teaching the course or certifying the learner's understanding.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: X · @RoupenMD
Frequently asked questions
How does the quiz track material already covered?
The poster says answering a question strikes through the corresponding part of the transcript.
Were the questions independently checked?
No independent accuracy check or measured learning improvement is established by the post.
More details
About this account
Claude Code generated an HTML question interface from the supplied transcript, including the requested link between answered questions and covered passages.
Primary post by @RoupenMD: https://x.com/RoupenMD/status/2114552077552877897. Post text and author bio checked through public FxTwitter syndication on 2026-10-26. Results reported by the poster, not independently reproduced. Suggested prompt by Agents In Use, not the author's original prompt. No excluded Nous Research affiliation appears in the retrieved bio.
Credit: @RoupenMD on X
Source published
Added to Agents In Use
Source checked
Tools and topics
Claude Code · CS50 lecture transcript · HTML
Study aids · Quizzes · Transcript coverage
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.