What It Actually Takes to Replace Five Chinese Learning Apps With One
Ask in any serious Chinese-learning forum which app to use, and nobody names one. The honest answer is always a stack. Anki for retention, because its scheduling has the most research behind it. Pleco for looking words up on the fly, because nothing else has its dictionary depth. Skritter if you actually want to write by hand, because typing pinyin teaches you nothing about stroke order. Du Chinese once you can read a paragraph, because its graded stories come with real native audio. Maybe a tutor on the side, because none of the above will ever tell you how you actually sound. That’s five logins and five habits to maintain before you’ve learned to say anything useful, and the reason nobody recommends fewer is that historically, no single app has done all five jobs well.
That’s the gap Merry Mandarin was built to close. What follows is an honest look at what closing it actually delivers, including the places it still falls short of a tool built for exactly one job.
What one app is actually standing in for
Each of those five apps earned its reputation doing one thing well, and it’s worth naming what that one thing is before claiming to replace it.
Anki’s reputation rests entirely on its scheduling algorithm, and as of version 23.10 in late 2023, that algorithm is FSRS, not the older SM-2 most flashcard apps still run. FSRS was built by a developer named Jarrett Ye, trained on roughly 700 million real Anki reviews from 20,000 users, and its own benchmarks show it needs 20 to 30 percent fewer reviews than SM-2 to hold the same retention. Merry Mandarin’s review queue runs on FSRS-5 as well, the same family of algorithm, already built into the review screen with no deck-building or manual configuration required.
Pleco’s reputation rests on its dictionary and its camera lookup, both genuinely excellent and worth using on their own. Merry Mandarin’s dictionary is its own, built entry by entry with context, nuance, and real example sentences rather than a bare definition swap, and it falls back to CC-CEDICT, the same open dataset much of the field draws from, for anything not yet covered. Every entry gets a tap-to-look-up card inside a reader or a course, and there’s a camera mode of its own: point it at a menu, a page, or a sign, and every word on it becomes tappable and checked against the dictionary. It’s word-level lookup, not full-sentence translation, which is a real and worth-naming limit rather than a gap to gloss over.
Skritter’s reputation rests on handwriting, specifically on strict stroke-order enforcement with instant correction. Merry Mandarin includes the same kind of writing practice, built on the same animated stroke-order approach, paired with character decomposition into the radicals and components a character is actually built from, so a new character arrives as a small, familiar set of parts rather than a fresh wall of strokes to memorize from nothing.
See a real stroke-order animation for any character, freeThe same underlying practice used in the app's writing exercises, no account needed
Try it yourselfNobody’s reputation covers pronunciation correction well, because it’s the hardest of the five to automate honestly. Merry Mandarin’s attempt is a speech-practice exercise that listens as you read a sentence aloud and grades your tones syllable by syllable, on the device, with nothing recorded or sent anywhere. It’s a grading system, not a conversation partner, and it will never replace the judgment of an actual tutor on the phrasing and naturalness a machine can’t hear.
Du Chinese’s reputation rests on graded reading with real native narration. Merry Mandarin’s version of that is a story and dialogue library that’s hand-written for the app rather than licensed from a stock bank, updated constantly rather than shipped once and left alone, with every line narrated aloud and every unfamiliar word tappable mid-story instead of requiring a separate lookup app running alongside it. A separate news reader sits next to it: real current-events articles adapted for learners, for the days you’d rather read about something that actually happened this week than a story written to teach a grammar point.
| What you need | The app most learners reach for | What Merry Mandarin does instead |
|---|---|---|
| Retaining vocabulary long-term | Anki, hand-built decks on SM-2 or FSRS | A built-in FSRS-5 review queue, no deck-building required |
| Looking up words while reading | Pleco's dictionary and camera lookup | Its own dictionary built for context and example sentences (CC-CEDICT as backup), tap-to-look-up in every reader, plus its own camera lookup |
| Practicing handwriting and stroke order | Skritter | Built-in stroke-order writing practice with character decomposition into radicals |
| Getting your tones corrected | A tutor, or nothing at all | Syllable-by-syllable tone grading as you speak, on the device, nothing recorded or sent anywhere |
| Reading graded stories with native-style audio | Du Chinese | A hand-written, constantly updated story and dialogue library, narrated line by line, words tappable mid-story |
| Reading about real current events | Nothing most course apps offer | A dedicated news reader with current articles adapted for learners |
| Making sense of your own recordings or podcasts | Nothing most learners have tried | A built-in transcriber that cleans up the audio, separates speakers, and transcribes it, on the device |
That last row has no real competitor to name, which is worth pausing on. Feed the transcriber a voice memo, a podcast clip, or a recording of an actual conversation, and it cleans up the background noise, works out roughly how many people are speaking and separates their turns, then transcribes the Chinese, all without the audio leaving the device. The free tier caps it at five minutes a month, enough to test it on something real before deciding whether it earns a place in your routine.
The honest pitch isn't that any one of these does its job better than a tool built for exactly that job. It's that you stop needing five separate logins to make progress on a Tuesday.
The engine underneath all of it
Every feature above eventually needs to be remembered, and that’s the one job none of the five specialist apps can outsource to something else in the stack. This is why the scheduling algorithm matters more than any single feature: a course with excellent content still fails a learner if the review timing is wrong, resurfacing words either so often they waste time or so rarely they’re forgotten before the next review. FSRS models forgetting directly, fitting a stability and difficulty estimate to every card from the reviews you’ve actually given it, rather than applying the same fixed multiplier to every card the way SM-2 does. We’ve written the full mechanical explanation, including the specific failure mode in SM-2 called ease hell, in a dedicated piece on why Merry Mandarin doesn’t use SM-2.
What the app actually tells you to do each day
None of the above matters if the daily habit doesn’t survive contact with a busy week, so the routine the app steers you toward is deliberately narrow. Onboarding asks you to pick a daily time commitment, not a vague goal, and shows a projected vocabulary outcome for each option so the number feels concrete rather than aspirational. Whatever number you pick, the onboarding tour ends on one explicit rule: clear your review queue before starting a new lesson, every day, in that order.
-
Clear the review queue first
FSRS has already worked out which cards are about to be forgotten. This is the highest-value five minutes of the day, before anything else.
-
Take the next lesson in the course
New tones, characters, grammar, or vocabulary, delivered in the app's own fixed sequence rather than a self-picked topic.
-
Spend anything left over on exposure
A story, a dialogue, a few minutes of the audio transcriber on something real. Optional, and never at the expense of the first two steps.
That ordering is not a small detail. A new lesson always feels like more visible progress than a review session, which is exactly the illusion behind why gamified apps built around lesson completion stall learners at the beginner wall once the easy streak-building runs out. Reviews first means the words you half-know get consolidated before more get piled on top of them. It’s a smaller, less exciting rule than most onboarding flows bother to state directly, which is probably why so few competing apps state it at all.
Where the other apps still win
None of this is a case for one app beating five specialists at their own games, and it would be dishonest to pretend otherwise. Pleco’s dictionary has been refined for over a decade and carries paid add-on dictionaries, like Oxford and Longman editions, that go deeper than any single built-in dataset. Skritter’s raw handwriting mode, with the snap-to-guide turned off entirely, is a sharper tool for someone specifically chasing beautiful, natural handwriting rather than correct stroke order. And nothing in Merry Mandarin replaces an actual tutor: there is no live conversation partner, no marketplace, no human on the other end correcting the thing only another person can hear. There’s also no public leaderboard or friend system, by choice rather than oversight, since the features that make a habit stick here are the review queue and the routine, not social pressure from other users.
So which is actually fastest to fluency
Fastest is a claim worth being precise about rather than just asserting. The retention engine is measurably more efficient per review than the algorithm most course apps and flashcard decks still run, which is a real, benchmarked claim, not a marketing one. But the bigger effect is quieter than an algorithm: a habit that only requires opening one app, in a fixed order, every day, survives more weeks than a habit that requires opening five apps in the right order and remembering which one does what. Research on habit formation puts the average time for a new daily habit to feel automatic at 66 days. The apps that get you to fluency fastest are, in the end, whichever ones you’re still actually opening on day 67, and a five-app stack loses far more learners to that arithmetic than any of the five ever loses to a weak algorithm.