thinkhard.SI

Prompt repository Research & learning

Turn a lecture or video transcript into notes you can study from

You have the transcript of a lecture or training video and want organized notes and self-test questions that stick to what the speaker said.

Registry no.
RES-035
Collection
Research & learning
Laid out for
Any chat model
Score
10 out of 10
Blanks to fill
5
Added

Research & learningAny chat

The prompt

Score 10 of 10, Excellent

Use it when You have the transcript of a lecture or training video and want organized notes and self-test questions that stick to what the speaker said.

  • {{learner}}
  • {{context}}
  • {{transcript}}
  • {{question_count}}
  • {{length}}
Full prompt and why it works
Prompt
Turn the transcript below into study notes for {{learner}}, so that they can review the material in a fraction of the time it took to watch, find any point again by its timestamp, and test whether they have really understood it. The notes are for learning this material, so they must say what the speaker said, not what a textbook might add.

What the material is (course, session, speaker's role) and what I am studying it for: {{context}}

Transcript, with timestamps if it has them:
"""
{{transcript}}
"""

Use these headings:
1. The main ideas: 5 to 8 sentences, one per idea, each with the timestamp where it starts.
2. Notes by section: follow the order of the talk, with a short heading for each part, bullets for the points, and the timestamp in brackets, for example [12:40].
3. Definitions: a table with the columns Term, Definition as the speaker gave it, Timestamp.
4. Examples the speaker used, one line each, with what each example was meant to show.
5. Test yourself: {{question_count}} questions that check understanding rather than recall, such as "Why does the speaker say the second method fails on small samples?" Put the answers, with timestamps, after all the questions.
6. Unclear in the transcript: words that look misheard, sentences that break off, and points the speaker said they would come back to but did not.

Keep the notes under {{length}}, not counting the questions. They should read clearly enough to study from a week later without the recording.

Use only the transcript. Do not add facts, definitions, formulas or examples from outside it, even to fill a gap, and do not correct the speaker silently; if a statement looks wrong or a term looks misheard, keep it and mark it [CHECK: why] so I can look it up. If the transcript has no timestamps, leave them out rather than estimating them. If a section seems to be missing, such as a sudden jump in topic or time, say so under heading 6.

Variables

{{learner}}
Who the notes are for, for example "me, studying for a certification exam".
{{context}}
The course or session, the speaker's role, and what you need from it.
{{transcript}}
The transcript, pasted in full, with timestamps if you have them.
{{question_count}}
How many self-test questions you want, for example 10.
{{length}}
The most the notes should run to, for example "1,200 words".

Why it works

  • What to avoid. Nothing is added from outside the transcript and nothing is corrected silently, so the notes match what will be tested on this material.
  • An example. Timestamps in a fixed bracket form tie every note and answer back to the moment in the recording.
  • If unsure. Misheard words, broken sentences and possible errors are kept and marked [CHECK] instead of being smoothed over.

How to use it

  1. Press Fill in blanks and type your details into each blank. What you type stays in this browser.
  2. Copy the filled prompt into the AI tool you already use, such as ChatGPT, Claude or Gemini, or open it in the Grader or the Prompt Tester here.
  3. Before you rely on it, try it on three or four real examples. The Prompt Tester runs them all at once and shows which ones go wrong.
Use it from code

Get this prompt as JSON from the free REST API, with your own text in the blanks, or let an AI assistant fetch it through the MCP server. No key.

curl -s https://thinkhard.si/api/v1/prompts/transcript-to-study-notes

With the MCP server added to Claude, ChatGPT, Cursor or VS Code, call get_prompt with {"prompt": "transcript-to-study-notes"}, or just ask for it by name.