Prompt Library

30 NotebookLM Flashcard Prompts for Anki and Drills

30 copy-paste prompts

Ask for tab-separated cards. Cover defs and applications. If it is not in the files, skip the card.

In short: This page contains 30 copy-paste ready prompts, organized into 6 categories with a description and pro tip for each. The first 5 prompts are free instantly, no signup needed. Hand-curated and tested by the AI Academy team.

By Louis Corneloup · Founder, Techpresso

Last updated September 10, 2026·Hand-curated & tested by the AI Academy team

Deck

5 prompts

Tab 25

1/30

✨ What it does

NotebookLM does make 25 tab-separated cards from [SECTION] from [UPLOADED FILES] for [JOB]. Paste your facts, then keep the draft if every number still matches your notes.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, make 25 tab-separated cards from [SECTION]. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Anki import.

Tab 50

2/30

✨ What it does

NotebookLM does make 50 cards across the whole notebook from [UPLOADED FILES] for [JOB]. Fill the brackets, then cut any line the model invented.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, make 50 cards across the whole notebook. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Tag source.

Defs

3/30

✨ What it does

NotebookLM does 20 definition cards, no yes/no from [UPLOADED FILES] for [JOB]. Run it once, then keep the take you can defend from the source.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, 20 definition cards, no yes/no. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Force a term.

Cause

4/30

✨ What it does

NotebookLM does 15 cause-effect cards from [UPLOADED FILES] for [JOB]. Paste the source first, then star the section you will actually use.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, 15 cause-effect cards. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: From files.

Apply

5/30

✨ What it does

NotebookLM does 10 application cards (what would you do) from [UPLOADED FILES] for [JOB]. Give it the real constraint, then drop anything it could not cite.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, 10 application cards (what would you do). Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: From cases in files.

Cloze

5 prompts

Cloze

6/30

✨ What it does

NotebookLM does 20 cloze deletions with one blank each from [UPLOADED FILES] for [JOB]. Paste your facts, then keep the draft if every number still matches your notes.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, 20 cloze deletions with one blank each. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Not two blanks.

Formula

7/30

✨ What it does

NotebookLM does cloze the formula name and the use from [UPLOADED FILES] for [JOB]. Fill the brackets, then cut any line the model invented.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, cloze the formula name and the use. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: If present.

Date

8/30

✨ What it does

NotebookLM does cloze dates only if printed from [UPLOADED FILES] for [JOB]. Run it once, then keep the take you can defend from the source.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, cloze dates only if printed. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Else skip.

Name

9/30

✨ What it does

NotebookLM does cloze people the files introduce from [UPLOADED FILES] for [JOB]. Paste the source first, then star the section you will actually use.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, cloze people the files introduce. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: No extra bios.

List

10/30

✨ What it does

NotebookLM does cloze a 3-item list the files give from [UPLOADED FILES] for [JOB]. Give it the real constraint, then drop anything it could not cite.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, cloze a 3-item list the files give. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Keep order.

Hard

5 prompts

Trap

11/30

✨ What it does

NotebookLM does 10 cards on distinctions the files say people mix up from [UPLOADED FILES] for [JOB]. Paste your facts, then keep the draft if every number still matches your notes.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, 10 cards on distinctions the files say people mix up. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Cite.

Weak

12/30

✨ What it does

NotebookLM does 10 cards on my weak list [WEAK LIST] from [UPLOADED FILES] for [JOB]. Fill the brackets, then cut any line the model invented.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, 10 cards on my weak list [WEAK LIST]. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: I paste it.

Oral

13/30

✨ What it does

NotebookLM does 8 cards I can answer out loud in 20 seconds from [UPLOADED FILES] for [JOB]. Run it once, then keep the take you can defend from the source.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, 8 cards I can answer out loud in 20 seconds. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Short answers.

Case

14/30

✨ What it does

NotebookLM does 6 cards from a case in the files from [UPLOADED FILES] for [JOB]. Paste the source first, then star the section you will actually use.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, 6 cards from a case in the files. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: No new case.

Num

15/30

✨ What it does

NotebookLM does cards for every number the files call must-know from [UPLOADED FILES] for [JOB]. Give it the real constraint, then drop anything it could not cite.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, cards for every number the files call must-know. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Skip fluff nums.

Like these prompts? There are full tutorials behind them.

Learn the workflows, not just the prompts. 300+ easy-to-follow tutorials inside AI Academy — and growing every week.

Try AI Academy Free

Export

5 prompts

Anki

16/30

✨ What it does

NotebookLM does remind field order: question, answer, source from [UPLOADED FILES] for [JOB]. Paste your facts, then keep the draft if every number still matches your notes.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, remind field order: question, answer, source. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Tab.

Quizlet

17/30

✨ What it does

NotebookLM does same cards as term/def lines from [UPLOADED FILES] for [JOB]. Fill the brackets, then cut any line the model invented.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, same cards as term/def lines. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Clean.

CSV

18/30

✨ What it does

NotebookLM does CSV with quotes escaped from [UPLOADED FILES] for [JOB]. Run it once, then keep the take you can defend from the source.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, CSV with quotes escaped. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: I import.

Reverse

20/30

✨ What it does

NotebookLM does also emit reverse cards for defs only from [UPLOADED FILES] for [JOB]. Give it the real constraint, then drop anything it could not cite.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, also emit reverse cards for defs only. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Not for cases.

Fix

5 prompts

No Yes

21/30

✨ What it does

NotebookLM does rewrite any yes/no card I paste from [UPLOADED FILES] for [JOB]. Paste your facts, then keep the draft if every number still matches your notes.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, rewrite any yes/no card I paste. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Force retrieval.

No Extra

22/30

✨ What it does

NotebookLM does drop cards not grounded in files from [UPLOADED FILES] for [JOB]. Fill the brackets, then cut any line the model invented.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, drop cards not grounded in files. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Delete them.

Level

23/30

✨ What it does

NotebookLM does match [LEVEL] from [UPLOADED FILES] for [JOB]. Run it once, then keep the take you can defend from the source.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, match [LEVEL]. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Do not raise.

Count

24/30

✨ What it does

NotebookLM does stop at [N] cards from [UPLOADED FILES] for [JOB]. Paste the source first, then star the section you will actually use.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, stop at [N] cards. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: I name N.

Redo

25/30

✨ What it does

NotebookLM does rebuild after I add notes from [UPLOADED FILES] for [JOB]. Give it the real constraint, then drop anything it could not cite.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, rebuild after I add notes. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Same tags.

Go from copy-pasting to actually mastering AI.

AI Academy: 300+ hands-on tutorials on ChatGPT, Claude, Midjourney, and 50+ other tools. New tutorials added every week.

Start Your Free Trial

Checks

5 prompts

Sources Only

26/30

✨ What it does

NotebookLM does answer [QUESTION] from [UPLOADED FILES] only from [UPLOADED FILES] for [JOB]. Paste your facts, then keep the draft if every number still matches your notes.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, answer [QUESTION] from [UPLOADED FILES] only. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Else NOT IN SOURCES.

Cite Check

27/30

✨ What it does

NotebookLM does give every claim a citation from the files from [UPLOADED FILES] for [JOB]. Fill the brackets, then cut any line the model invented.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, give every claim a citation from the files. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Or drop.

Gap Check

28/30

✨ What it does

NotebookLM does list what the files lack for [JOB] from [UPLOADED FILES] for [JOB]. Run it once, then keep the take you can defend from the source.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, list what the files lack for [JOB]. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Homework.

Format Check

29/30

✨ What it does

NotebookLM does output Markdown I can paste from [UPLOADED FILES] for [JOB]. Paste the source first, then star the section you will actually use.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, output Markdown I can paste. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Clean.

Stop Check

30/30

✨ What it does

NotebookLM does the deliverable with no pep after it from [UPLOADED FILES] for [JOB]. Give it the real constraint, then drop anything it could not cite.

<context> Sources: [UPLOADED FILES] Job: [JOB] </context> <task> Using ONLY the uploaded sources, the deliverable with no pep after it. Cite the source for each claim. If the files are silent, write NOT IN SOURCES. Do not add outside knowledge. </task>

💡

Pro tip: Stop.

Free tool

Prompt Optimizer

Turn a rough idea into a structured, professional AI prompt.

Try it free →

Frequently Asked Questions

These prompts tell it to stay inside uploaded files for the flashcards. If a fact is missing it must say so. Still click the citation chips.

PDFs, docs, slides, or pasted notes that actually cover the job. Empty notebooks produce empty answers.

Some prompts steer the overview. Others write study guides or briefs in the chat. Load sources before you run either.

Yes. NotebookLM itself is a Google product with its own limits. Check your account before a long session.

The hub mixes jobs. This page is only the flashcards: thirty prompts so the notebook stays on one deliverable.

Prompts are the starting line. Tutorials are the finish.

A growing library of 300+ hands-on tutorials on ChatGPT, Claude, Midjourney, and 50+ AI tools. New tutorials added every week.

Start 7-Day Free Trial Explore AI for Consultants

7-day free trial. Cancel anytime.

Browse All Prompt Collections

Popular prompt collections

AI Photo Editing Prompts ChatGPT Jailbreak Prompt Best ChatGPT Prompts Funny Hinge Prompts ChatGPT Caricature Prompts

30 NotebookLM Flashcard Prompts for Anki and Drills

Deck

Tab 25

Tab 50

Defs

Cause

Apply

Cloze

Cloze

Formula

Date

Name

List

Hard

Trap

Weak

Oral

Case

Num

Export

Anki

Quizlet

CSV

Tags

Reverse

Fix

No Yes

No Extra

Level

Count

Redo

Checks

Sources Only

Cite Check

Gap Check

Format Check

Stop Check

Frequently Asked Questions

Prompts are the starting line. Tutorials are the finish.