Start here
Paste three things into an ordinary AI chat that accepts text: (1) the lecture transcript, (2) the course or unit name (optional), (3) any topics you want emphasized (optional). Then paste the full prompt below the transcript.
If the transcript is very long, the prompt will ask you to split it by section. Do that instead of trimming it silently.
How to check the output
- Every card back should contain a phrase you can find in your transcript. Search the transcript for it.
- Cards tagged
[gap]should say "Needs source..." instead of an invented answer. - Administrative lines (deadlines, room changes) should be skipped and reported as skipped.
- Card count should be 3-8. If higher, the prompt asks you which concepts to merge.
What this prompt does
It converts a raw transcript into question-first spaced-repetition cards, each with a front question, a short back answer drawn only from the transcript, a transcript anchor, a confidence tag ([clear], [partial], [gap]), and a one-line note on why the concept matters for the unit. It also produces a 7-day review plan and a "Verify before studying" list.
Prerequisites and permissions
No account setup, terminal or API is needed. You only need the transcript text and a chat window. The prompt works from text you supply in the conversation; it does not open files, browse, or read course platforms on its own. It does not enroll you, submit work, save files, or contact anyone, and it does not guarantee exam results.
Limitations
- Answers are limited to what your transcript states. A missing detail becomes a
[gap]card, not a guess. - Transcripts with unclear audio, mixed languages or heavy filler produce weaker cards.
- The review schedule is a draft you adjust, not a fixed study prescription.
- Always verify cards against your instructor's slides or assigned reading.
FAQ
Do I need to upload the lecture file? No. Copy the transcript text into the chat. If you have no transcript, this prompt cannot work.
What do I do with the [gap] cards?
Ask your instructor or check the assigned reading, then rewrite the card yourself once you have the answer.
Verification note
Source reviewed; runtime not tested. The prompt text is the original TokRepo prompt (CC BY 4.0). A reference link to ChatGPT release notes is context only and is not proof of any specific provider feature.
Complete reusable prompt
Role: You are a study-methods assistant helping a student convert a raw lecture transcript into spaced-repetition cards. You work only from text the student supplies in this conversation. You cannot open files, browse the web, or read course platforms on your own.
Input I will paste below:
- The lecture transcript (or a section of it).
- The course or unit name (optional).
- Any specific topics I want emphasized (optional).
If any required input is missing, incomplete, or unreadable, stop and list exactly what you need before proceeding. If the transcript is too long for one pass, tell me and ask me to split it by section.
Task: Step 1. List the 3-8 main concepts actually stated in the transcript. Do not add outside facts, definitions, or examples. Step 2. Write one spaced-repetition card per concept. Each card must have:
- Card number and concept label
- Front (question-first, framed as a prompt to retrieve, not a yes/no question)
- Back (answer written only from transcript content, in one or two sentences)
- Transcript anchor (the phrase or line in the transcript the answer comes from)
- Confidence tag: [clear] if the transcript states it plainly, [partial] if the transcript hints but does not fully state it, [gap] if the concept is mentioned but the transcript gives no usable answer Step 3. For every [gap] card, write "Needs source: ask instructor or check assigned reading" instead of inventing an answer. Step 4. Add one short note per card on why that concept matters for the unit, drawn from the transcript only. Step 5. Create a 7-day review plan: which card numbers to review on day 1, day 2, day 4, and day 7, based on the confidence tags. Never schedule a [gap] card as if it were answerable.
Output format:
- Header: unit name, transcript section, total card count, count of clear/partial/gap cards.
- Table or numbered list of cards in the five-part structure above.
- Review plan table with day number and card numbers.
- A short "Verify before studying" list naming every card the student should check against the actual lecture slides or textbook.
Handling uncertainty:
- If a concept could be interpreted two ways, show both readings and mark the card [partial].
- If the transcript contradicts itself, show both versions side by side and mark [partial].
- If a section is mostly administrative talk (deadlines, room changes), skip it and say so.
Before any action outside this chat: do not claim to send these cards anywhere, save files, or contact anyone. You are producing a draft the student will copy into their own flashcard app or notes. State this limitation in the output header.
Review checks I will run:
- Does every back answer contain a phrase I can find in the transcript? If not, the card fails.
- Does any card use facts not in the transcript? If yes, delete or mark it [gap].
- Are the front questions answerable without seeing the back? If a question is vague, rewrite it.
- Does the review plan avoid treating [gap] cards as study-ready?
- Is the card count between 3 and 8? If higher, ask me which concepts to merge.
Tone: plain, direct, no filler. Do not praise the lecture. Do not add motivational language. Keep cards short enough to fit a standard flashcard app.
Fictional example input: Unit: Biology 101, Lecture 4 on cell transport. Transcript excerpt: "Today we covered diffusion, which is movement of particles from high to low concentration. Osmosis is a special case involving water across a membrane. Active transport goes against the gradient and needs energy. Don't forget the lab report is due Friday and the room changed to B204."
Illustrative output shape: Concept: Diffusion. Front: "What is diffusion and which direction do particles move?" Back: "Movement of particles from high to low concentration." Anchor: "diffusion, which is movement of particles from high to low concentration." Confidence: [clear]. Why it matters: foundation for comparing other transport types. Concept: Energy source for active transport. Front: "What does active transport require that diffusion does not?" Back: [gap] - transcript mentions energy but not the specific molecule. Needs source. Confidence: [gap]. Note: administrative lines about the lab report and room change are skipped as non-conceptual.
Pass/fail checks for this example:
- Pass if the diffusion card's back quote matches the transcript phrase exactly.
- Pass if the active-transport card is marked [gap] rather than filled with an invented molecule name.
- Fail if the room change appears as a study card.
- Fail if any card claims a fact such as "active transport uses ATP" that the excerpt does not state.
Boundaries: This prompt prepares a study draft only. It does not enroll you, submit work, or guarantee exam results. Always verify against your instructor's materials.
References and reuse
- ChatGPT release notes · Reviewed 2026-10-04
Original TokRepo prompt · CC BY 4.0. Reference documents retain their own rights.