NotebookLM Flashcards: Good for Understanding, Not for Remembering
TL;DR: NotebookLM can now generate flashcards from your sources, and they are genuinely good for a first pass at new material. But NotebookLM does not schedule those cards with spaced repetition, so it helps you understand a topic without making sure you still remember it weeks later. For that you need an app that generates cards and then schedules them, like Imprimo, which pairs AI card creation with the FSRS algorithm. Use NotebookLM to understand, use a spaced repetition app to remember.
Google added flashcards and quizzes to NotebookLM in early 2026, and the feature is good enough that a lot of students stopped reaching for a separate card maker. You upload your sources, ask for flashcards, and get a tidy set drawn straight from your own material.
I use NotebookLM and I like it. But after a few weeks of leaning on its flashcards, I hit the same wall every time. The cards are great for getting my head around a topic, and then they just sit there. Nothing brings them back. That gap is the whole point of this post.
What NotebookLM flashcards actually do
NotebookLM is a research and understanding tool. You give it sources, a stack of PDFs, your notes, an article, and it can summarize them, answer questions about them, and now turn them into flashcards and quizzes.
The flashcards are grounded in your sources, which is the best thing about them. They do not hallucinate facts from outside your material, so a card about your pharmacology lecture stays about your pharmacology lecture. You can edit the cards and share a set with a classmate. For a first pass at a new chapter, this is a fast way to check whether you understood what you read.
If your goal is "help me make sense of these sources today," NotebookLM does that well.
Where it stops short
Here is the catch. NotebookLM generates the cards and then leaves the rest to you.
There is no spaced repetition. The cards do not get scheduled. There is no daily due queue, no tracking of which cards you keep missing, and no algorithm deciding when to show a card again so it lands right before you would have forgotten it. You flip through the set, feel good, and close the tab.
That feeling is the trap. Recognizing an answer when you see it is not the same as being able to recall it cold three weeks later, which is what an exam asks for. The whole reason spaced repetition exists is that memory fades on a predictable curve, and the only reliable fix is to retrieve the fact again at widening intervals. A static set of cards you reviewed once does almost nothing against the forgetting curve.
NotebookLM is also built for the desktop and web, where you do research. Daily review is a phone habit. It happens in the four minutes you have in a queue or between classes, and a research tool open in a browser tab is not where that habit lives.
So you end up with generation without scheduling, which solves half the problem. The making of cards is the easy half. The remembering is the hard half, and that is the half NotebookLM does not touch.
NotebookLM flashcards vs a spaced repetition app
| NotebookLM flashcards | Spaced repetition app (Anki, Imprimo) | |
|---|---|---|
| Generates cards from your sources | Yes | Yes |
| Grounded in your own material | Yes | Yes |
| Schedules reviews over time | No | Yes |
| Tracks the forgetting curve per card | No | Yes |
| Daily due queue | No | Yes |
| Built for phone-first daily review | No | Yes |
| Best at | Understanding new material | Long-term retention |
The two tools are not really competing. One helps you understand a set of sources. The other makes sure you still know it months later. The mistake is expecting NotebookLM to do the second job, then wondering why the facts did not stick.
How to close the gap
You do not have to choose. The simplest workflow keeps both tools doing what each is good at.
Use NotebookLM the way it is designed: to read, summarize, and understand your sources. Ask it questions, get your bearings on a hard chapter, let it pull out the main points.
Then take the same source into a spaced repetition app for the part that lasts. You already have the PDF, the notes, or the link. Feed it to an app that generates cards and then schedules them.
This is exactly what Imprimo is built for. You upload the PDF, paste the text, or share a URL, it drafts a deck in seconds, you trim it down to the cards worth keeping, and then FSRS, the modern spaced repetition algorithm, schedules each card based on how well you actually remembered it. You get NotebookLM's fast generation and the review timing it leaves out, in one app that lives on your phone where daily review actually happens.
If you want the longer argument for why generation alone is not enough, do AI flashcards actually work covers what separates a useful AI card from a useless one, and the AI flashcard maker guide walks through how to get cards worth reviewing.
A note on AI-generated cards in general
Whether the cards come from NotebookLM, Imprimo, or anywhere else, the same rule applies: review the draft before you trust it. AI is good at drafting cards and bad at knowing which facts matter to you. The fastest path to a clean deck is to let the AI do the typing, then delete the cards that test trivia and keep the ones that test the thing you actually need to recall.
A good app makes that editing step part of the flow rather than an afterthought. In Imprimo the draft review gate sits between generation and your deck for this reason, so trimming is a couple of taps, not a chore. Most flashcards are bad explains what to keep and what to cut.
The short version
NotebookLM flashcards are a real upgrade for understanding your sources, and they are worth using for that. They are not a study system, because they stop at generation and skip the scheduling that turns a card into a memory.
- Use NotebookLM to understand new material.
- Take the same sources into a spaced repetition app for retention.
- Always review the AI draft before you trust the deck.
If you want one app that generates cards from your PDFs and notes and then schedules them with FSRS so they stick, Imprimo is free to start on the App Store. Drop in the same source you would give NotebookLM and see the difference a real review queue makes.
For more on the algorithm doing the scheduling, spaced repetition apps explained covers what these tools do and whether you need one.
