What Anki is
Anki is a flashcard program built on spaced repetition. You write cards with a prompt on one side and an answer on the other; Anki shows each card at expanding intervals — a day, then three, then a week, then a month — timed so that you review each fact just before you would have forgotten it, and rates your recall to adjust the next interval.
The scheduling is not a productivity heuristic. The spacing effect and testing effect are among the most replicated findings in cognitive psychology, with over a century of supporting literature, and Anki is a direct implementation of both. That is a rare position for a tool: its central claim is not a marketing assertion but a body of research you can go and read. Medical students discovered this and built the practice into their culture; most researchers have not, and quietly relearn the same terminology every year.
Why researchers use it
- The mechanism is evidence-backed — you can check the literature behind it rather than taking a vendor’s word.
- It ends the relearning cycle — statistical definitions, your field’s terminology, the key claim of a seminal paper, kept available instead of re-looked-up every six months.
- Writing the card is half the work — condensing a concept into a single question is a comprehension test disguised as note-taking, and it fails informatively when you do not understand something.
- Entirely yours — open source, local decks, free sync, and export to plain text. No subscription and no lock-in.
- Language learning — for anyone doing fieldwork or reading literature in a second language, it is the standard tool for a reason.
Where it fits in a research workflow
Anki holds the small set of things worth having available instantly rather than looked up. That set is smaller than enthusiasts suggest and larger than most researchers assume.
Concretely: the terminology of a field you are entering, the assumptions of the statistical tests you use weekly, the handful of formulas you should not have to check, the key finding and sample size of the ten papers your work responds to, and — for talks — the numbers you need to answer questions without reaching for a slide. It sits alongside Obsidian or Logseq rather than replacing them: notes hold your understanding and your synthesis, Anki holds the retrievable facts. When a literature note contains a definition you will need repeatedly, that definition becomes a card; the rest of the note stays where it is.
Getting started
Twenty minutes to start, and the daily habit is the entire product.
- Install desktop Anki and set up free AnkiWeb sync. The Android client is free; the iOS one is the project’s single paid item and funds development.
- Make your own cards rather than downloading a shared deck. Writing the card is where the learning happens, and someone else’s deck teaches you their understanding of the material.
- Keep cards atomic — one fact, one card. A card asking three things is a card you will fail for the wrong reasons and eventually delete.
- The step people skip, and the one that decides everything: review daily, briefly. Ten minutes every morning works; ninety minutes on Sunday does not, because the scheduling assumes you show up. A deck reviewed sporadically becomes a backlog and then an abandonment.
Anki vs the alternatives
| Alternative | Does it better | Pick it if |
|---|---|---|
| Obsidian | Holding understanding, synthesis, and connections | The material is an argument, not a fact |
| Zotero | Keeping the papers themselves findable | You need the source, not the recall |
| Quizlet | A friendlier interface and ready-made decks | You want convenience over control and scheduling quality |
| Re-reading | Nothing — it reliably produces familiarity, not recall | Genuinely nothing; this is the illusion Anki exists to break |
Anki is not competing with note tools. It does a different job, and the two combine well.
Cost, licensing, and your data
Free and open source under AGPL for desktop, Android, and the web client, with the iOS app as a one-off paid purchase that funds the project. Sync through AnkiWeb is free.
Decks are local files with a documented format and a full export path, so nothing is locked in. The consideration worth naming is what you put in a card: cards synced through AnkiWeb are stored on the project’s servers, so a card containing an unpublished result, a participant identifier, or anything under a data agreement is a small disclosure that most people would never think to make deliberately. Sync can be disabled, or a self-hosted sync server used, if that matters for your material.
The honest review
Strengths. Anki works, it is free, and its central claim is backed by a literature rather than a testimonial. Over a career the compounding effect is real: the terminology, formulas, and findings you deliberately retain stop being things you look up and become things you think with, which changes how fast you can read a new paper in your own field.
Limitations. The interface is utilitarian to the point of unfriendliness, and the advanced features — note types, card templates, the add-on ecosystem — have a genuine learning curve that is entirely disproportionate to their benefit for most users. It requires daily discipline and punishes lapses with a backlog that feels punitive. And the fundamental constraint: it builds recall, not understanding. A card can make you fluent in a definition you could not apply.
Verdict. Adopt it if you are entering a new field, learning a language, or repeatedly relearning the same terminology — those three cases have clear payoffs. Skip it if your bottleneck is synthesis, writing, or thinking, where it will feel productive and be nothing of the kind. The condition that flips the answer is whether your problem is remembering or understanding.
Common questions
Is Anki free?
Yes on desktop, Android, and the web, under an open-source licence, with free sync. The iOS app is a one-off paid purchase that funds the project’s development.
What should researchers actually put in Anki?
Things you need instantly and repeatedly: field terminology, the assumptions of tests you use weekly, key formulas, the central finding and sample size of the papers your work responds to, and vocabulary for a second language. Not: anything you can look up in thirty seconds and rarely need.
How long does daily review take?
Ten to fifteen minutes for a deck of a few hundred mature cards. The scheduling assumes daily attendance — reviewing sporadically produces a backlog that grows faster than you can clear it, which is the usual reason people abandon it.