Anki alternatives on iOS, honestly compared

It's 11pm on a Tuesday. The lesson ran twenty minutes long, my tutor is offline, and there are 14 words on a notepad in front of me. One of them is madrugada, which I asked about because she used it and I had no idea what it meant, and which I won't be able to define by Thursday.

Making a card out of it means typing a sentence, typing a translation, finding audio, picking an image. Call it 30 seconds. Times 14, times every lesson, forever.

I did not enter those words. The page is still in the notebook. I know what it says without opening it, which is somehow worse.

That's the actual reason people go looking for an Anki alternative on iOS. The search usually starts one step later, at the App Store listing for AnkiMobile, where the Get button should be and $24.99 is instead.

The case for not switching

Anki on desktop is free. AnkiDroid is free. The sync service is free. So are the decks other people have already built for Spanish and French, and some of them are very good.

It runs FSRS, has an add-on ecosystem nothing else comes close to, and your review history is yours, in a file, exportable forever.

The $24.99 is one payment. At $19.99 a month our own cheaper plan passes it in under seven weeks and keeps going. I'd rather say that than hope you don't do the arithmetic.

Anki's own listing explains where the money goes, and the explanation is fair:

The listing describes AnkiMobile as written by the lead developer of Anki and AnkiWeb, and the program itself as free, multi-platform and open-source. One paid app funds the free ones everywhere else. Calling that a paywall misses what it's paying for.

So here's the concession, up front, where it costs me something. If you have a deck you like and you'll keep feeding it, buy AnkiMobile and close this tab. Nothing beats a tuned Anki deck belonging to someone who enjoys tuning it, and no app I could build will change that.

Keep reading only if that sentence describes a person you've been failing to become.

The switch nobody flips

Here's what half this search results page gets wrong. FSRS is off by default. The Anki manual is unambiguous: FSRS is "an alternative to Anki's legacy SuperMemo 2 (SM-2) algorithm," and you "enable FSRS under the 'FSRS' section, at the bottom of the deck options page." Bottom of the page. Global toggle.

So a lot of people running Anki are on the legacy SM-2 scheduler while assuming they're on the modern one, because a blog told them FSRS became the default and it didn't.

A second wrinkle matters more if you're new. FSRS gets good by fitting itself to your history, analyzing "your review history" to "generate parameters that are optimal for your memory," and the manual lists a "Low number of reviews (less than a few hundred)" as a reason it underperforms. So the algorithm everyone recommends Anki for is the thing Anki can't give a beginner on day one.

Worth being precise about how good FSRS is, since every competing page waves at it in one clause. The open benchmark scores schedulers across 9,999 collections and 349,923,850 reviews. FSRS-7 posts a log loss around 0.3437. A research model called RWKV-P beats it at 0.2773, and nobody ships RWKV-P in anything you can install.

srs-benchmark, read 22 Aug 2026

0

collections in the evaluation set

0M

reviews, excluding same-day repeats

0

appearances of SM-2 in that table

Sit with that gap. A neural model scores better than FSRS on a third of a billion real reviews, and you can't install it. So FSRS is the best scheduler that ships in something you can download and tunes itself to a real person. That claim is defensible. The stronger version people write is not.

The real reason people leave

Building cards turns out to be a job. It ends more Anki habits than the algorithm and the $24.99 combined.

Migaku publishes the honest number, in a guide arguing for making your own cards, which is why I trust it: "If you spend 30 seconds per card (Typing front, back, maybe adding an image)." For a 1,500-card deck, "that's 750 minutes or about 12.5 hours."

Twelve and a half hours, at a pace assuming you never stop to check a gender or rewrite a sentence. Thirty seconds is optimistic, and it's their optimism, not mine.

And the failure has a shape. You do it for a week, diligently. Then one lesson runs long, you promise yourself you'll enter the words tomorrow, and that page is still untouched in March. The cards you built keep coming up; the words your tutor watched you grope for never entered at all. That gap matters more than how many cards a day you do.

None of which makes cards the whole method. If you're doing zero reading and zero listening, fix that before you fix your app. But it's why "alternative to Anki" is a strange query taken literally. Anki's scheduler is not the part that failed you.

The other iOS apps, and who each one suits

You searched for alternatives, plural, and so far I've named one app and my own. Here's the rest of the shortlist, and my method so you can discount it: App Store listings pulled from Apple's own lookup API on 22 August 2026, every capability claim taken from what the developer publishes about themselves. No controlled trial, and I'm the last person you should take as neutral. Read the who-it-suits line and ignore my adjectives.

Wokabulary, version 8.1.0, is the page currently ranking first for this query, and a better product than that fact made me want to admit. It's a real native app across iPhone, iPad, Mac and Watch, it syncs through your own iCloud account rather than someone's server, and it's free up to 60 words with no signup and no time limit. It does one job: holding vocabulary you collected yourself. If you want a tidy Apple-native word list you own outright and you're happy typing the entries, it beats Verbamor on platform coverage and on being free to try. I'd send a Mac-first learner there before I'd send them to me.

Two caveats. Past 60 words its own listing says "a subscription is needed," the same model its landing page dings Anki for, and the figure appears nowhere. And its scheduling is words sorting into learning levels, asked more or less often. That's a spaced repetition system, and nothing there claims FSRS.

Mochi, version 26.7.1, free to download, is markdown notes and flashcards in one place, with cards made from notes you already wrote. It suits anyone who takes structured notes anyway and resents copying them into cards a second time.

Space, version 5.3.0, also free to download, builds cards from a PDF, a URL or a photo of your notes. If your source material is a document, that's the shortest path from document to deck here.

Space is the only one of the three that takes your Anki history with you. Its App Store listing says it imports “Anki (with your learning progress intact), CSV, Excel, Google Sheets”, and its 5.0.0 notes say the import “now preserves your learning progress”. I have not tested it, so that is the developer’s claim rather than mine, and it is the single most useful line on this page for anyone with a deck they care about.

Verbamor has no Anki import at all. Moving here means starting the deck fresh from your next lesson, and if the history is what you are protecting, Space deserves the look before I do.

And notice what the shortlist is sorted by. Every app here competes over where the cards come from: your typing, your notes, your documents, or your speech.

Anki's own warning about apps like mine

The AnkiMobile listing has a line aimed squarely at products like mine, and I'd rather quote it than pretend it isn't there:

Two of the three charges land on us. We have far fewer features than Anki. We charge a subscription, and it costs more than $24.99 once. Only the name part misses.

My defense is that a feature list only pays out on cards that exist. Anki has every feature I lack and it still lost to my notepad. You get to decide whether that describes you, and there's six months of evidence in a notebook.

What you're actually buying

It's Thursday now. Madrugada came up again, and I asked what it meant a second time. My tutor was kind about it. She typed it into the chat window with a little sentence underneath, the way you'd write something down for a child, and I felt the specific heat of having wasted both our time on a word I'd already been given.

That word cost me two lessons. The vocabulary I failed to record was the most expensive thing in the room, and the only thing I could have kept for nothing.

So: Verbamor is $19.99 a month, or $129.99 a year with a 7-day trial on the annual plan only. No free tier. iOS only, five languages, and it cannot import your Anki deck. If you switch, you start empty and rebuild from your next lesson forward.

What it sells is Migaku's 30 seconds. You say madrugada into it, and the sentence, the translation, the native audio and the image come back built, scheduled by FSRS.

The tempting sentence is that the typing goes to zero. It doesn't, quite. You still say or type 1,500 words to get 1,500 cards. It beats typing a sentence, hunting a translation and finding an image by hand, but I don't have a measured figure to put next to Migaku's 30 seconds, and I'm not going to invent one to sit beside a number a competitor actually published. You still have to open the app on the night the lesson ran long, the exact habit Anki failed to get out of me.

Faster, not free, against 12.5 hours of typing. I'd rather publish that honestly than claim the zero.

I built it because I was the man with the untouched notebook, and I'd spent long enough trying to be the other man to stop believing I'd become him. That's the trade: money for the 12.5 hours, and less control than Anki gives you.

And if you go with Anki, which for a lot of you is the right call, do the one thing this search results page forgot: open deck options, scroll to the bottom, turn FSRS on. It's a 30-second job most of the internet told you was already done.

---

Sources

Every load-bearing claim Verbamor makes is traced to its paper on the research page.