Memory

Is Anki good for vocabulary, and which settings work?

Yes, for keeping the words you choose, as long as you review almost every day. Anki brings each card back shortly before you are likely to forget it, and in a semester-long study of 62 beginning Spanish students, more days of Anki use went with better end-of-semester results. For vocabulary: turn on FSRS, keep desired retention at the default 90%, start with 10 to 15 new cards a day rather than the default 20, and keep learning steps shorter than a day. Anki's manual puts the cost at roughly 200 reviews a day for 20 new cards a day.

By the Readlang Mobile team · Updated 31 October 2026

Is Anki good for vocabulary?

Yes, for remembering words you have already met. Anki is a flashcard program that schedules reviews with spaced repetition: each card returns after a gap that grows as you keep recalling it. Spaced review is one of the most consistent findings in memory research, including studies with foreign-language words; our guide to spaced repetition summarises that evidence.

Direct evidence for Anki itself is thinner. Seibert Hanson and Brown (2020) followed 62 university students in a semester-long beginning Spanish course who were assigned to study with Anki as a course requirement. Students who used it on more days did better on end-of-semester measures of Spanish, even after the authors controlled for starting ability, motivation, self-efficacy and beliefs about learning. The title of the paper, "an effective but bitter pill?", names the catch: the students were reluctant to use the app and reported low enjoyment.

Anki has two limits worth knowing before you start:

• It does not choose the words. A card helps you keep a word; whether the word deserves a card depends on how often you will meet it. The research on how many words you need suggests learning the most frequent 2,000–3,000 word families deliberately and meeting the rest through reading and listening.
• It works only if you review. Cards you skip pile up. Most of the settings below are about keeping the daily review load small enough that you keep going.

Which Anki settings work for vocabulary?

Most of Anki's defaults are fine for vocabulary. The ones worth changing are the number of new cards, the review limit and, once FSRS is on, the learning steps. The table lists the default of each setting, what we suggest for vocabulary, and the source. Settings are in the deck options, opened from the gear icon next to a deck.

After changing settings, do not press "Reschedule cards on change" on your first switch from the old scheduler: the manual says it is not recommended then, because many cards may become due at once (Anki manual).

What desired retention should you use for vocabulary?

Leave it at 90% unless you have a reason to move it. Desired retention is the chance FSRS aims for that you still remember a card when it comes up for review. "Higher retention leads to shorter intervals and more reviews per day", and the default of 90% "offers a good balance of retention and workload" (Anki manual).

• Higher. "Above 90% the workload increases very quickly, and above 97% the workload can be overwhelming." Going from 90% to 95% for a vocabulary deck buys a few points of recall at the cost of many more reviews.
• Lower. At lower desired retention you do significantly fewer reviews a day but forget more cards (Anki manual). For vocabulary you also meet in reading, 85% is a reasonable trade, because reading tops up the words between reviews.
• Range. The FSRS tutorial calls 80–95% reasonable and says "90% should work fine for most people" (FSRS tutorial).

To see the trade on your own collection, use the simulator in the FSRS section: it estimates your workload in reviews or minutes a day for the settings you choose (Anki manual). The older "Compute minimum recommended retention" button was removed in Anki 25.07, so guides that tell you to press it are out of date.

How many new cards per day should you set in Anki?

Start with 10 to 15 and adjust by the time reviews take. There is no research figure for the right number; the limit comes from the reviews each new card brings later. Anki's manual gives the one rule of thumb, roughly 200 reviews a day for 20 new cards a day, and warns that "more than one Anki user has excitedly studied hundreds of new cards over their first few days of using the program" and then become overwhelmed (Anki manual).

Scaling that ratio gives a rough picture of what each setting means over a year:

Three things change the picture:

• Reverse cards double the count. With a note type that makes a card in each direction, 10 new words a day are 20 new cards and about 200 reviews. Start with recognition cards (foreign word, then meaning) and add the reverse only for words you want to use.
• Time, not count, is the real budget. Anki's statistics show your average time per card; multiply it by the review count. If reviews take longer than you can keep up every day, lower new cards first, desired retention second.
• Cards are not word families. One card per word form or meaning means 2,000 cards cover fewer than 2,000 of the word families the vocabulary research counts.

Our spaced repetition guide explains why the review load keeps rising for weeks after you start, and the first-week plan sets the limit up together with a daily reading routine.

Should you switch Anki to FSRS?

Yes, if your Anki is version 23.10 or newer. FSRS (Free Spaced Repetition Scheduler) is "an alternative to Anki's legacy SuperMemo 2 (SM-2) algorithm" and is supported in Anki 23.10, AnkiMobile 23.10, AnkiWeb and AnkiDroid 2.17+ (Anki manual). Instead of multiplying each interval by an ease factor, it models how memory of each card fades and schedules the review so that your chance of recalling it matches your desired retention (Anki FAQ; Anki manual).

• Optimize the parameters. The "Optimize" button fits FSRS to your own review history. Since Anki 24.06 there is no minimum number of reviews (FSRS tutorial), but the manual lists a low number of reviews, "less than a few hundred", among the reasons FSRS may perform poorly. Optimizing "once every month is sufficient" (Anki manual).
• Old habits in the history. If you used to press the answer buttons differently, "Ignore cards reviewed before" leaves those reviews out of the optimization (Anki manual).
• Grade honestly. FSRS learns from your answers. Press Again when you could not recall the word before turning the card, however familiar the answer looks afterwards; the hub explains why recall matters.

What should a vocabulary card in Anki look like?

A word with the sentence you met it in, and the meaning on the back. A card that shows context is easier to recall and closer to how you will meet the word again; our guide to learning words from reading covers why context and repetition work together. Keep each card to one word and one meaning, so that a failed card tells FSRS exactly what you forgot.

Make cards for words that recur. In our counts of 117 books, about half of the different word forms in a typical English book appear only once; a card for a word you may never meet again costs reviews for months.

Is Anki better than Quizlet for vocabulary?

For long-term review, Anki gives more control. Quizlet also offers spaced repetition: its App Store listing says Quizlet "intelligently resurfaces exactly what you need to review, right when you need it" (App Store). Anki schedules each card with FSRS on every platform it runs on and lets you set desired retention, limits and steps. Quizlet is quicker to start with shared sets and games; Anki suits a deck you plan to keep for years.

Clozemaster or Anki?

They do different jobs. Clozemaster gives you ready-made sentences with one word missing, so you practise words in context without making cards. Its reviews follow fixed default intervals of 1, 10, 30 and 180 days, and a wrong answer at any point sends the sentence back to 0% mastered; Pro subscribers can change the intervals (Clozemaster FAQ). Anki works with your own words and sets each card's interval from your history. The two can be combined: Anki for the words you meet in your own reading, Clozemaster for extra sentences.

SuperMemo or Anki?

SuperMemo is where Anki's original algorithm came from: Anki's legacy scheduler is based on SM-2 (Anki FAQ), the algorithm Piotr Woźniak wrote for the first SuperMemo program in 1987 (Woźniak, 1990). Today supermemo.com sells ready-made language courses, "25 languages and nearly 300 courses", with a free first month, scheduled by what it calls "a proprietary algorithm" (SuperMemo). Choose SuperMemo if you want a course to follow; choose Anki if you want to review your own words with a free, open scheduler.

When is Anki the wrong tool?

• When you cannot review most days. Missed days leave the reviews waiting, and at 10 new cards a day a week off can leave several hundred due cards. If daily review is unrealistic, lower new cards to five or pause them until the backlog is gone.
• For rare words. Beyond the first few thousand word families, there are too many words, each too rare, to card them all. Reading and listening at a comfortable level meets them often enough (learning words from reading).
• As the only practice. A card trains recognition or recall of a meaning. Using the word in speech and writing takes practice that Anki does not give.

References

1. Anki Manual. Deck options: daily limits, learning steps, FSRS, desired retention, simulator. docs.ankiweb.net/deck-options.html, checked 31 October 2026.
2. Anki FAQs. What spaced repetition algorithm does Anki use? docs.ankiweb.net, checked 31 October 2026.
3. Anki source code, default deck configuration (new_per_day 20, reviews_per_day 200, learn_steps 1 and 10 minutes, relearn_steps 10 minutes, desired_retention 0.9, maximum_review_interval 36,500 days, leech_threshold 8, leech_action tag only, bury_new and bury_reviews off). github.com/ankitects/anki, checked 31 October 2026; current version 26.09.3 per apps.ankiweb.net.
4. Open Spaced Repetition. FSRS4Anki tutorial (no longer updated; the Anki manual is the current reference). github.com/open-spaced-repetition/fsrs4anki, checked 31 October 2026.
5. Seibert Hanson, A. E., & Brown, C. M. (2020). Enhancing L2 learning through a mobile assisted spaced-repetition tool: An effective but bitter pill? Computer Assisted Language Learning, 33(1–2), 133–155. doi:10.1080/09588221.2018.1552975
6. Apple App Store (US). Quizlet: More than Flashcards, Quizlet Inc. apps.apple.com, checked 31 October 2026.
7. Clozemaster. FAQ. clozemaster.com/faq, checked 31 October 2026.
8. Woźniak, P. A. (1990). Application of a computer to improve the results obtained in working with the SuperMemo method. supermemo.com, checked 31 October 2026.
9. SuperMemo. Home page. supermemo.com, checked 31 October 2026.
10. Apple App Store (US). AnkiMobile Flashcards, Anki Software LLC. apps.apple.com, price checked 23 October 2026.