Start with 55 small quiz banks, then fold them into one big app: how a small Korean studio plays the exam-prep game

In the English store it's called "DayPrep: JLPT, TOPIK, HSK," which looks like a Japanese/Korean/Chinese language test app. In the Korean store the same bundle is listed as "데이프렙 - 자격증 시험 CBT 문제은행 모의고사" (professional license CBT question bank mock exams), and its installers are basically Korean office workers studying for certifications. A month after launch, it has only 1,409 installs on Google Play, so it hasn't broken out. The interesting part is the approach behind it: launch dozens of "one exam, one app" apps to capture store search, then merge them into a single app with AI-generated questions and explanations. This piece covers how it works, how much volume and money there is, who's competing, and whether the playbook is worth copying in other countries.

A busy office worker who doesn't want to memorize

On Nov 1, a Korean user left a five-star App Store review titled 「진짜 미쳤어요 돈 하나도 안아까움」 (seriously amazing, not a penny wasted). She said she first wrote it on her iPad and it didn't show up, so she wrote it again on her phone: "I used to carry heavy textbooks around and ask GPT about concepts I didn't understand every day. Then I found this. It has all the past exam questions and concept explanations too... I'm a working exam taker, and I don't have the time or energy to sit down and study."

The day before, Oct 31, was the 37th Korean Real Estate Agent licensing exam (per the Q-Net announcement). The app's Google Play installs went from 10-31 on that day to 339, then 713,11-02 and 1,290 (Google Play installs). Whether the two events are causally linked is unclear. Play's install counts also refresh only every few days (inferred). But the feature it launched that day did first open up Real Estate Agent, Real Estate Appraiser and Security Guidance Instructor exams in Korea (1.4.0 release notes, 11-01).

Its focus is Korean domestic certifications, not JLPT.

What it is

It's a question-bank exam prep app. The dayprep.co homepage reads 「60개+ 지원 시험、19만+ 문항」 (supports 60+ exams, 190k+ questions). Coverage is mainly Korean national licenses, technician certifications, the Korean Vocational Skills certificates, civil service exams and NCS (the national job competency test), plus JLPT N1–N5, TOPIK I/II and HSK level 6.

Each day you get a free allowance, and when it runs out you buy a pass. You can buy a single exam (1 months, 3 months, 12 months, or lifetime) or the 「자유이용권」 (free-use pass), which covers all exams. On Play, in-app purchases run from $2.19 to $74.99.

The features are detailed: each day it auto-builds a set mixing review material and weak chapters, uses spaced repetition, shows a chapter-level accuracy heatmap, explains why each option is right or wrong, and runs mock exams matching real question counts and timing, scored against the passing line. The Korean store description opens with the pitch: 「준비하는 시험이 여러 개여도 앱을 여러 개 깔 필요가 없습니다」 (even if you're preparing for several exams, you don't need to install several apps).

That line is a bit ironic, since the developer itself has dozens of apps.

Where the 55 mini apps came from

The developer is DayLab (주)데이랩 in Gangnam-gu, Seoul; the iOS developer name is dasong sim. From April to August 2026, it launched about 54 apps on iOS, and roughly 40 across two Play developer accounts (App Store developer page, AppBrain, Play developer page). Each app covers one exam: 정보처리기사 (information processing engineer), 요양보호사 (caregiver), 조리기능사 (cooking technician), HSK 6, JLPT N2……

The mini apps' titles are strings of search terms, like 「위험물기능사 필기 CBT 모의고사 문제은행 시험일정」 (hazardous materials technician written CBT mock exam question bank exam schedule). Each one has tiny volume: TOPIK II 50+ installs, HSK 6 10+, driving theory test 10+; on iOS each one has 0 to 44 ratings.

After October's all-in-one app launched, the terms pages for these small apps were all hung under dayprep.co/<exam name>/, and some titles even got a "- DayPrep" suffix, like "영양사 국가고시 문제은행 모의고사 - DayPrep" (Nutritionist National Exam Question Bank Mock Test). It looks like they're moving users from the small apps into the big one. How exactly the migration works, such as whether there are pop-ups inside the app, isn't visible from outside (unverified).

The output behind this most likely comes from AI. The store description reads "연습 문제와 해설은 공개된 시험 범위와 학습 목표를 바탕으로 데이랩이 자체 제작했습니다" (Practice questions and explanations were made in-house by DayLab based on publicly available exam scopes and study goals). A small team producing 60+ exams and 190k questions in half a year, plus 12 interface languages, would be hard without AI (inference).

Volume and money: people do pay, but not many

It has only appeared on the Korean charts. Japan, Vietnam and China, where JLPT and HSK users are concentrated, have never shown it. The English titles are just hanging there, and the volume isn't coming from them.

People are paying, and it's real. On the App Store in 10-17, a user wrote 「무료로 쓰다가 좋아서 프리미엄 결제했는데 후회안함」 (used it free, liked it, bought premium, no regrets). On Play in 10-27, another wrote 「유료결제 할까 말까 고민중」 (still deciding whether to pay). But Korean education apps get onto the top-grossing chart with a low bar, around #90, so it may only be a few in-app purchases a day (estimated). That shows someone is paying. It doesn't show the scale.

One more detail. One iOS review from 10-18 says: "기존에 다른앱쓰다가 오류가 너무많아서……광고하길래 설치해봣는데" (I was using another app with too many errors... I installed it because I saw an ad). So it probably ran ads, but it doesn't show up in the public Meta or Google ad libraries (those libraries not showing it doesn't mean it didn't run ads). The other channels visible are all small: Instagram @dayprep_app with 115 posts and 54 followers; YouTube Shorts with #데이프렙 clips; and an official site with a free exam calendar page organized by exam, used to capture search traffic. No media coverage, and nobody discusses it on X.

Who's competing

The Korean professional certification question bank market is already crowded.

The English-title line is even more crowded. Vietnam's Migii uses one app per exam: the JLPT version has 2.35M Play installs, the HSK version 800k, and the TOPIK version 610k; HSK Online SuperTest has 2.66M. DayPrep is three orders of magnitude smaller than them on this line, and it only covers HSK levels 6.

Its strength is one app holding dozens of exams, one pass covering all of them, plus daily question assignments. Reviews show users buying in: "원하는 시험은 다른앱에 없었는데 업로드도 빠르고 여러개 진행 할 수 있는것도 너무 좋고" (The exam I wanted wasn't in other apps, new exams go up fast, and I can work on several at once, which is great). Some users also push for new exams in reviews: "품질경영기사 꼭 출시주세요!!" (Please release the Quality Management Engineer exam!!), 11-04.

This is exactly where AI question banks work best: big companies skip niche exams, established question bank sites update slowly, and a small team can ship one in a few days.

If it were you

Verdict: B. Copying DayPrep itself isn't worth it. The Korean niche is already full. What you can take is a method: use AI to batch-produce question banks for niche exams, first split them into dozens of single-exam small apps to capture store search and see which exams people pay for, then consolidate into one all-in-one app sold as a pass. The method could be tried in another country, but it hasn't proven it can scale, and question quality and copyright are real risks.

Who it fits: indie developers or two-to-three-person teams who can build a question bank engine, produce questions and explanations reliably with AI, and understand a particular country's certification system (or have a local partner). Ideally someone who has taken these exams and knows where candidates get stuck.

Who it doesn't fit: anyone trying to go head-on against large language exams like JLPT, HSK, TOPIK or TOEFL, where players like Migii have millions of installs. Also not for anyone who can't proofread questions: certification candidates have zero tolerance for wrong questions, and a single bad question means a one-star review.

How to do it: pick a market with a strong certification culture, many niche national certifications, and rough local question bank apps. Japan's various 資格試験, and civil service exams in Vietnam, Indonesia and Brazil, are worth browsing in the stores first (inference, competition not verified one by one). Use AI to draft questions from public exam outlines, then have someone who actually took each exam review them. List the small apps with long titles like "exam name + question bank + mock test + past papers", and after three months see which ones have paying users. Then launch the all-in-one app and use the all-exam pass to lift average revenue per user. The same package can use different store titles in different languages, so one package captures search in two markets. On the website, offer a free exam calendar and an introduction page for each exam to catch web search.

How to acquire users: store search is the main channel, and exam names are the keywords. The peak is one to two months before the exam, so launch timing should line up with just before or after registration deadlines. Some reviewers installed after seeing an ad, which suggests small paid spends also work during exam season (inference).

When to move: there's no rush, since this approach doesn't depend on a single window. But each exam runs only once or twice a year, so missing one exam season means waiting six months, which is why the first batch of apps should go live before registration for the target exam opens.

Signs that would prove the read wrong: ① After three months DayPrep's Play installs are still in the thousands and it drops out of the top 200 on the iOS Top Grossing chart. That would mean "build a matrix, then consolidate" doesn't pull in many users even in a crowded market like Korea, so moving elsewhere needs more caution. ② It gets hit with copyright claims from exam bodies or publishers, or wrong-question complaints in reviews pile up. That would mean the risk of mass AI-generated questions is bigger than it looks. ③ Conversely, if within two or three months it breaks into the top 30 of the Korean education Top Grossing chart and also shows up in the Japan or Vietnam stores, the method can scale, and the B rating is too low, so it's worth moving faster.

Data and sources

• Install counts and per-country chart data: Google Play, 2026-10-29 to 11-05. iOS ratings count, charts and reviews: App Store Korean storefront, 10-28 to 11-06.
• Store description, content declarations, 1.4.0 release notes, Play reviews: Google Play English and Korean pages, 11-06. Website figures: dayprep.co, 11-06.
• Number of apps under the developer: App Store developer page, Play developer page, AppBrain.
• Real estate agent exam dates: Q-Net, announcement for round 37.
• Competitor scale: Google Play install counts and App Store rating counts, 11-06.
• Places marked "est.", "inferred", or "unverified" are estimates and inferences.