FSRS (Free Spaced Repetition Scheduler) is Anki's modern scheduling algorithm. It estimates how likely you are to remember each card and schedules the review for when that drops to your target. To use it, open Deck Options, switch on FSRS at the bottom, keep desired retention at 90%, and click Optimize. In the open srs-benchmark, FSRS predicted recall far more accurately than Anki's old SM-2 scheduler; see our spaced repetition statistics for the numbers.
Sources: Anki manual (deck options), open-spaced-repetition srs-benchmark
FSRS is the most significant improvement to Anki in years and aligns with the latest spaced repetition research trends. Here's what it is, why it's better than SM-2, how to enable it, and what to set your retention target to.
FSRS (Free Spaced Repetition Scheduler) is a new scheduling algorithm built into Anki that replaces the old SM-2 algorithm. It's based on the science behind spaced repetition intervals and a modern memory model, and it learns from your own review history, a good example of adaptive study.
FSRS calibrates from your review history, so the more well-formed cards you give it, the better it gets. If making cards is what's holding you back, StudyCards turns a PDF or your notes into an .apkg you import into Anki the normal way.
Exports a native .apkg. No credit card.
FSRS is based on the DSR (Difficulty, Stability, Retrievability) model of memory. Three variables govern every card:
How intrinsically hard is this card? Some cards are just harder to remember than others - different content, different connections, different personal relevance. FSRS tracks this per-card rather than applying a one-size-fits-all ease factor.
How long until you forget this card? Stability increases each time you successfully recall it, which is essentially how retrieval and storage strength work. A card you've reviewed 10 times has much higher stability than one you've reviewed twice. FSRS uses this to schedule reviews at the optimal moment.
What is the probability you can recall this card right now? This decays over time following a forgetting curve, a process that is central to optimizing memory decay with AI systems. FSRS shows you the card when R drops to your target retention level (e.g., 90%). That's the optimal review moment - not too early (wasting a review), not too late (after forgetting).
Requires Anki 23.10 or later. Update Anki first if needed.
| Retention | Reviews per day | Best for |
|---|---|---|
| 0.70 | Very low | Language learning, low-stakes content |
| 0.80 | Low | General studying, moderate importance |
| 0.90 | Moderate (recommended default) | Most learners, good balance |
| 0.95 | High | Medical school boards, high-stakes exams |
| 0.97+ | Very high | Not recommended: far more reviews for a marginal gain |
Note: the Anki manual says 90% "offers a good balance of retention and workload", that above 90% the workload increases very quickly, and that above 97% it can be overwhelming. Start at 90% and only raise it for a specific high-stakes exam.
The FSRS4Anki Helper add-on (code: 759844606) adds a full control panel for FSRS management. It lets you:
Related: Anki settings complete guide, spaced repetition schedule explained, what is spaced repetition.
FSRS stands for Free Spaced Repetition Scheduler. It is an open-source scheduling algorithm built into Anki that models how likely you are to remember each card and schedules the review for when that probability drops to your target.
For most people, yes. FSRS predicts recall far more accurately than Anki's old SM-2 scheduler in the open srs-benchmark, and it lets you choose a retention target. It works best once you have a few hundred reviews for it to learn from.
Open any deck's options, scroll to the FSRS section at the bottom and switch FSRS on. It applies to all presets at once. Then set desired retention and click Optimize so FSRS fits its parameters to your review history.
Keep desired retention at the default 90% unless you have a reason to change it, and click Optimize every month or so. The Anki manual warns that workload rises very quickly above 90% and can be overwhelming above 97%.
It can, but no single figure applies to everyone. The published srs-benchmark measures how accurately FSRS predicts recall, not how many reviews you save. Your workload depends mostly on your desired retention and how many new cards you add.
Unlike the old SM-2 algorithm, which used static formulas, FSRS is a machine learning model. When you click "Optimize," Anki reads your review history (every "Again," "Hard," "Good" and "Easy" you've ever pressed) and works out the weights that fit your memory. The algorithm then follows how you forget, instead of a generic average of all users.
To get the most out of it, keep these tips in mind:
FSRS lets you choose your "Desired Retention." Many new users set it too high. In spaced repetition, there is a non-linear relationship between the percentage of memory retained and the amount of time spent reviewing. This is known as the efficiency frontier.
As you push your retention target higher, the workload increases exponentially. For example, moving from 85% to 90% retention might add a manageable number of reviews to your day. However, moving from 95% to 97% adds a lot of reviews for a marginal 2% gain in memory. That is usually a poor use of your time.
FSRS is a scheduling tool. It decides when you see a card, but it cannot fix a badly written one. If a card is too wordy, ambiguous, or contains too many ideas, you will find yourself hitting "Again" regardless of how advanced the algorithm is. This creates "leech" cards that waste your time and skew your FSRS data.
To get the most from FSRS, pair it with atomic cards: break information into the smallest pieces you can. This keeps the "Difficulty" value in the DSR model stable and predictable.
FSRS gets more accurate the more cards it has to learn from. The bottleneck for most people isn't reviewing, it's getting cards into Anki in the first place. StudyCards reads a PDF or your notes and gives you an .apkg so FSRS has something to work with.
Exports a native .apkg. No credit card.
Generate Anki flashcards from PDFs