The best spaced repetition apps in 2026, sorted by who writes the cards

Every app in this comparison does the same core thing: it shows you a question, hides the answer, asks how the recall went, and decides when to show you the question again. The scheduling is a solved problem. The best algorithms are open source and most of the serious tools use one of two or three of them. If you are choosing between spaced repetition apps on the strength of the scheduler, you are choosing on the thing that varies least.

What varies is everything around the schedule, and three questions cover most of it. Who writes the cards. What happens when you miss a week. Whether you can take your cards and your history with you when you leave. The rest of this article is eight tools answered on those three questions, a table, and a verdict by the kind of learner you are, as of September 2026.

One disclosure before the list: Agent Memo is the tool we make. It is in the comparison because leaving it out would be stranger than including it, and it is described in the same terms as the others. Weigh the paragraph accordingly.

The three questions

Who writes the cards. In the classical tools, you do, one at a time, and this is both their strength and the reason most people stop using them within two months. Writing a good card is real work; a good card asks for recall rather than recognition, carries the reason and not just the fact, and can be traced back to where it came from. Newer tools draft cards for you from notes, PDFs, or highlights. The question then becomes how many they draft, how good they are, and whether you can see where each one came from.

What a missed week costs. Reviews are due daily by design. Miss a week and the tool has a choice: show you the whole pile, or fit what is due into a session you will actually do. This decides whether you are still using the app in month three, and it is the question the marketing pages never answer.

Whether you can leave. Plain-text export, and, if you are coming from Anki, an import that keeps your scheduling history instead of resetting every card to new. A tool without export is a tool you cannot leave without starting over, which is a strange thing to accept from something you are trusting with years of reviews.

The tools

Anki

The reference point, and for good reason. Free and open source on desktop, Windows, Mac and Linux; free on Android; a one-time purchase of about $25 for the iPhone app, which funds the rest. The largest library of shared decks in existence, an add-on ecosystem that can make it do almost anything, and since late 2023 a modern scheduler as an option alongside the classical one.

You write every card. There is no drafting from a source. Shared decks cover the big syllabuses, medicine and languages especially, but a deck someone else wrote is a deck about what they wanted to remember.

A missed week is shown in full. Come back to four hundred due cards and Anki shows you four hundred due cards. There are settings and add-ons to soften this; the default is the pile.

Export is complete, in its own format and as plain text, and most other tools import it. If you want total control and are willing to do the authoring, nothing else comes close. If the authoring is what has stopped you before, nothing about Anki has changed that.

SuperMemo

Where the algorithms came from. Piotr Wozniak's program has been running since the 1980s and its scheduler is the ancestor of the one in Anki. It is Windows-first, paid, and has an interface that has not changed with fashion.

Its distinctive feature is incremental reading: you import whole articles and the program schedules the reading itself, extracting passages into cards as you go. For a certain kind of reader this is the most complete system ever built. For most it is more system than they will use.

You write the cards, with the program helping you extract them. Missed weeks are handled by a postpone feature that is explicit about what it is doing. Export exists and is less commonly supported by other tools than Anki's.

Mochi

A calm, markdown-native spaced repetition app for people who found Anki ugly and Quizlet childish. Cards are markdown; decks are files; the interface gets out of the way. Free with limits, paid for sync across devices.

You write the cards. Missed weeks are shown as due but the daily presentation is gentler than Anki's. Export is plain markdown, which is as portable as export gets. Mochi has no capture story: nothing reads a source and proposes cards. It is the best-looking tool in the list for someone who already likes writing cards.

RemNote

A notes app first, with spaced repetition built into the outliner: a line in your notes can be marked as a card and scheduled. Free tier, paid for the rest. Since 2023 it has added a generator that drafts cards from PDFs and from your notes, which is where most of its recent growth has come from.

Cards are written by you inside your notes, or drafted by the generator in quantity. The generator's instinct is volume; a chapter can come back as sixty cards, and choosing which to keep is on you. The scheduler is modern.

A missed week is shown. Export covers the notes and the cards. If your notes already live in an outliner and you want the cards to live beside them, this is the tool; if you do not take notes that way, the notes app is in the way of the cards.

Quizlet

The largest study platform by users, built for students with an exam next month. Free with a paid tier that has moved features behind it more than once; check the current free tier before assuming. Its Learn mode spaces items within a study session, and a generator drafts sets from your notes.

Cards are written by you, drafted by the generator, or found in the enormous library of shared sets. Long-term scheduling across weeks and months, of the kind that keeps a fact for a year, is not what the product is optimised for; it is optimised for the test on Friday.

Missed weeks are not really a concept; there is no daily due queue in the classical sense. Export is limited compared with the tools above. For a student, it may be the right choice. For an adult trying to keep what they read, it is a different product.

Readwise

The tool for people who highlight. Readwise collects highlights from Kindle, Kobo, Apple Books, read-later apps and a long list of other sources, and resurfaces a few every day. Its Mastery mode, added more recently, turns highlights you choose into active recall prompts scheduled by a recall-probability model. Subscription only.

The cards are your highlights. You chose them while reading, one at a time, and the prompt is the highlight itself rather than a question written for recall, which means the retrieval is often "what was this passage" rather than "what did it argue". Mastery is opt-in per highlight, so the selection work stays with you.

Missed weeks are handled gently; the daily review is a fixed small number rather than a growing pile. Export is good. If you already read on a Kindle and already highlight, Readwise is the lowest-friction path to reviewing what you marked. It does not write questions and it does not choose for you. A fuller comparison covers the tools around it.

Recall

A save-everything tool with review attached. Recall captures articles, videos, podcasts and PDFs, summarises them, connects them in a graph, and lets you chat with the collection. It then offers spaced repetition over the saved summaries, with generated quizzes. Freemium, with a large user base.

The cards are drafted by the tool from its own summaries, which is one step further from the source than a card written from the passage; checking a card against what the author actually said means going back through the summary. The primary job of the product is retrieval, finding the thing again, and the review mode is a feature of that rather than the point.

Missed weeks accumulate; a large library means a long queue. Export covers the saved content. If what you want is a searchable second brain with a review mode, this is a serious option. If what you want is to need the second brain less, the priorities are the other way round.

Agent Memo

The tool we make, described in the same terms. You capture a source: paste a link, drop a PDF or an EPUB, a YouTube video, a recording, a spreadsheet, a note. The agent reads the whole source and proposes the cards worth keeping, a dozen from a book rather than four hundred, each traceable to the passage it came from. You see all of them before any enter review; dropping one is a click; you can write your own beside them.

The day has a ceiling you set, five, ten or twenty minutes, and the schedule fits inside it. A missed week comes back as a fuller session, not a wall. The scheduler is modern. Import from Anki keeps your review history; export is plain text and an Anki-compatible format. There are no streaks and no badges, on purpose.

What it does not do: decide what you wanted from the book. The agent proposes from the source; the question of what mattered to you is still yours, and the reading method it fits into starts before the capture. It also does not offer the shared-deck library Anki has or the notes app RemNote is. It is a review tool for adults who read, and it is narrow by design.

Side by side

Prices and features as of September 2026. Every one of these changes its tiers more often than articles get updated; check before you decide.

For reading, provenance matters as much as the review schedule: you should be able to check a card against its source.

A verdict by learner

You have a syllabus and an exam. Anki, with a shared deck for your subject if one exists, or Quizlet if the exam is soon and the material is factual. Both are built for this and both are free enough.

You take notes in an outliner and want cards beside them. RemNote.

You enjoy writing cards and want a tool that respects that. Mochi, or Anki if you also want the add-ons.

You read on a Kindle and already highlight. Readwise, with Mastery turned on for the highlights that matter.

You want to save everything and find it again later. Recall.

You read books, articles, papers and videos, you have tried to make cards from them and stopped, and you want the reviewing without the authoring. Agent Memo. This is the person we built it for, and the description of the others above is honest about who they serve better.

Frequently asked questions

Is Anki still the best spaced repetition app? For control, price and community, yes, as of 2026. For someone who has tried it and stopped because of the authoring or the backlog, the answer is that Anki has not changed on either point, and the tools that address them are the ones to look at.

Do these apps use the same algorithm? Broadly. Anki, RemNote and Agent Memo all use modern schedulers descended from the same research; Readwise uses a recall-probability model of its own; Quizlet's Learn mode spaces within a session rather than across months. The scheduler is the least important difference between them.

Can I move from Anki to another app without losing my progress? Depends on the app. Agent Memo imports the scheduling history; several others import the cards but reset them to new. Check for "keeps review history" specifically, not just "imports Anki".

Are AI-generated cards any good? Some are, and the way to tell is whether the card can be traced to what the author actually said. A card drafted from a summary of a summary cannot be checked. A card that cites its passage can. More on this.

What about apps for languages specifically? Most of the tools here handle vocabulary well; Anki's shared decks are strongest for it. Dedicated language apps add pronunciation and grammar drills that a general spaced repetition tool does not, and are a separate comparison.

What to do now

Answer the three questions for yourself before you pick: do you want to write the cards, what will you do after a missed week, and will you want to leave one day. Then take the tool that matches, put a dozen cards from the last thing you read into it, and see whether the session is still happening in two weeks. If the authoring is the step you know you will skip, start with the source and let the agent propose the cards.