AI Flashcards & Spaced Repetition

A memory palace gets something into your head vividly. Spaced repetition keeps it there. It is the quiet workhorse behind almost everyone who remembers a lot on purpose: medical students, language learners, memory champions. The idea is almost insultingly simple, the science behind it is overwhelming, and AI has just removed its single biggest chore. Here is how it works and which tools to use.

What is spaced repetition?

Back in 1885, Hermann Ebbinghaus measured how fast he forgot things and drew the result: the “forgetting curve,” a steep slide where most of what you learn leaks away within days. But he also found the fix. Each time you review something just as it is about to fade, the curve flattens, and the memory lasts longer before the next review. Spaced repetition is simply that, done deliberately: revisit material at growing intervals (a day, then three, then a week, then a month) so each review lands right before you would have forgotten. You do less work and remember more.

Why does it work so well?

Because two of the most reliable findings in all of memory research stack on top of each other here. The first is the spacing effect: a landmark review by Cepeda and colleagues pooled 184 studies and found spaced practice beats cramming, every time, across ages and subjects. The second is retrieval practice: work by Karpicke and Roediger, published in Science, showed that the effort of pulling an answer out of your head (rather than re-reading it) produces far stronger long-term memory. A good flashcard review does both at once. You retrieve the answer, and you do it on a spaced schedule. That combination is about as close to a proven study method as cognitive science has.

What is FSRS, and why does it matter?

The catch with spaced repetition is timing: review too early and you waste effort, too late and you have already forgotten. For decades the popular app Anki used a fixed formula (SM-2) to guess the timing. In 2023 it switched its default to FSRS, the Free Spaced Repetition Scheduler, and it is a real upgrade. FSRS is a machine-learning model that studies your own review history to predict when you personally will forget each card, then schedules the review for just before that moment. You set a target (say, “remember 90% of my cards”), and across analyses of hundreds of millions of reviews it hits that target with roughly 20 to 30 percent fewer reviews than the old formula. In plain terms: the AI under the hood learns how you forget, so you study less and remember the same amount.

How does AI make spaced repetition easier?

The reason most people quit Anki is not the reviewing. It is the card-making. Turning a textbook chapter into good flashcards by hand can take hours, and it is tedious. That chore is exactly what modern AI tools now handle. Point them at a PDF, a YouTube lecture, a web page, or your own notes, and they draft a deck of question-and-answer cards in seconds, then schedule them for you. The barrier that stopped most people from ever starting is gone.

Which AI flashcard tools should you use?

Tool	What it does	Cost
Anki	The gold standard. Endlessly customizable, now runs FSRS. You (or AI) make the cards	Free (paid iOS app)
StudyGlen	Generates cards from PDFs, text, images, and YouTube; uses FSRS	Free tier, then paid
RemNote	Note-taking and flashcards in one; write notes, get cards, review with spacing	Free tier, then paid
The Sponge	Turns any web page into a scheduled deck with one click	Paid (trial)
Quizlet	Huge library of ready-made sets; AI features behind a subscription	Free tier, then paid

For most people, the best combination is Anki for the engine (it has the strongest algorithm and runs on everything) and an AI generator to make the cards. Use an assistant like Claude to draft cards from your notes, or a dedicated tool to pull them from a PDF. More options live in our AI tools directory.

How do you build a system that sticks?

Most people are well served by one app off the shelf, and that is more than enough. But AI now lets you go further and assemble a custom setup that used to need a developer. My own is a good illustration. I wire three tools together: my Anki flashcards, my Obsidian notes, and Slack as the control panel, with AI doing the building, wiring, and fine-tuning while I actually use the system day to day. When a note lands in Obsidian, the pipeline can turn its key facts into Anki cards; I run and tweak the whole thing from Slack; and when a piece needs changing, I have AI rebuild that part of the plumbing. The specific tools matter less than the point: a non-engineer can now build and maintain a personal memory system, with AI as the developer. Whatever you use, the principle is the same: capture in notes, convert to cards, and let spaced repetition run the schedule.

Is AI-generated study as good as making your own cards?

Partly, and it is worth being clear about the trade. Writing a flashcard yourself forces you to decide what matters and put it in your own words, and that effort encodes the idea before you ever review it. A deck the AI hands you skips that step. But the heavy lifting of spaced repetition happens at review time, when you retrieve the answer, and that part is identical no matter who made the card. The sensible approach: let AI draft the cards so you actually start, then edit them into your own words as you go. You get the best of both, the same way a memory palace works best when you still do the mental work rather than letting the tool do all of it.

Common questions

What is spaced repetition?

Reviewing material at growing intervals so each review lands just before you would forget. It flattens the forgetting curve, so you remember more with less total study.

Is Anki still the best?

For the algorithm and flexibility, yes. Anki now uses FSRS, the strongest scheduler available. Pair it with an AI tool to make the cards and you remove its only real downside.

What is FSRS?

The Free Spaced Repetition Scheduler, Anki’s default since 2023. It learns your personal forgetting pattern and schedules reviews to hit a retention target with 20 to 30 percent fewer reviews than the old formula.

Can AI make my flashcards for me?

Yes. Tools like StudyGlen, RemNote, and The Sponge generate cards from PDFs, videos, web pages, or your notes. You can also have Claude or ChatGPT draft them.

Should I make cards myself or let AI do it?

Both. Let AI draft so you actually start, then reword cards into your own language as you review. The review itself, where the memory forms, is the same either way.

Sources

• Anki: what spaced repetition algorithm it uses (SM-2 to FSRS)
• FSRS (Free Spaced Repetition Scheduler), open-source project
• Cepeda et al. (2006), meta-analysis of the spacing effect; Karpicke & Roediger (2008), Science, on retrieval practice.

Last reviewed: June 2026. The spacing and retrieval-practice findings are long established; AI flashcard tools and pricing change often, so confirm current details on each tool’s site.

You may also like

• Build a Memory Palace with AI
• Best AI for Studying (2026)
• AI for College Study
• Why Handwriting Still Beats Typing for Learning
• How to Use Claude AI
• The daily Beginners in AI newsletter