
A ₩60k four-week pass: a Korean TOEIC Speaking AI mock-test app, TOAST, squeezed into the Top Grossing chart on 3,800 installs
What it is An AI mock-test app for Korean TOEIC Speaking test-takers. A free 10-minute diagnostic gives an estimated score, then you buy passes lasting 24 hours to 4 weeks, with unlimited mock tests, templates, and pronunciation correction during the pass.
Why it took off The category leader 토스미 had already shown that test-takers will pay for short unlimited mock-test access before an exam. TOAST followed at about 30% less, and launched with the playbook its team used on a previous vocabulary app: sponsored Instagram reviews, good ratings stacked at launch, and a name that borrows from "토스". It landed in the second-half hiring season.
BHow you could follow Don't build this in Korea, which is already crowded. Pick an exam with an expensive registration fee, fixed test dates, and a speaking or writing section that self-study can't crack (for example, the Japanese Eiken second-round interview or Taiwan's GEPT speaking test). Copy the "free diagnostic + short pre-exam pass" model, and make the question bank and scoring more solid than TOAST's.
Est. monthly revenue (Google Play) under $1KWe estimated this from its rank on country top-grossing charts and real install counts, calibrated against real revenue; typical error is 2–3x. Read it as a tier, not an exact number.
- Data (updated daily) · launched 2026-08-18 · 11-19 installs 4,267 · launched 2026-08-18 · Korea rating 500 · charted in 11-20 of 1 stores
Android · Total installs worldwide Data 2026-11-09 → 2026-11-19 (11 days, 8 observations). Google's detail page only refreshes install counts every other day, so a flat day doesn't mean nobody installed.iOS · best category chart rank · by store Data 2026-10-28 → 2026-11-20 (24 days, 24 observations). Lower rank is higher; gaps mean it wasn't charted that dayiOS · rating count · Korea Data 2026-10-28 → 2026-11-20 (24 days, 24 observations). Apple doesn't publish download counts; rating growth is a proxy for downloads.
A 9-person indie team in Korea, August, built an AI-scored TOEIC Speaking mock-test app (Koreans call it "토스"). It doesn't sell monthly subscriptions. It sells passes lasting "24 hours to 4 weeks, unlimited during the pass," priced from ₩12,000 to ₩62,000. On Google Play it has only 3,790 installs, yet in Nov 9 it hit #38 on the Korean Education Top Grossing chart. Most of the money comes from iPhone: on the App Store Korean Education Top Grossing chart it has sat between #22 and #40 every day since late October. It isn't a hit, and the product isn't better than the leader's. What's worth studying is the playbook: find an exam with an expensive fee and fixed dates, then sell "unlimited mock tests" to people desperate for a score in the last one or two weeks.
"₩60k a month... hmm"
In 2026 Oct 7, someone left a review of TOAST on the Korean App Store. The title was "앱 자체는 좋음" (the app itself is good), and the body was one line: "앱 자체는 좋은데 돈 내야되네 몰랏음 한달에 6만원..흠……." (The app itself is good, but I had to pay, I didn't know. 60 ten-thousand won a month... hmm...)
They still gave it five stars.
This review gets the business right. 60k won (about $43, at 1 dollars ≈ 1,450 won) is absurdly expensive for an app, but a TOEIC Speaking test costs ₩84,000 to register (namu.wiki), and most test-takers are there for jobs: 60% of 2025 test-takers cited job hunting, and ages 25–29 make up 52.1% (한국TOEIC위원회, reported by newstap). If the score falls short, you pay the registration fee again, so next to that, a four-week pass doesn't look expensive.
What it sells
Open the app and take a free 10-minute diagnostic. The AI scores your estimated level against each part of the TOEIC Speaking test and flags your weak spots. Enter your target grade (IM3, IH, AL…) and exam date, and it schedules what to practice each day. Everything after that is paid: 10 full timed mock tests with 11 questions each, matching the real exam; 200 answer templates for each question type; and real-time pronunciation and intonation correction while you shadow-read. The store listing is blunt about who it's for: "Job seekers self-studying for IH and AL" and "People with 1–2 weeks left before the exam who need to cram."
Here's how the passes are priced, next to the category leader 토스미 (TOSME) (App Store Korea pricing page):
| Duration | TOAST | 토스미 |
|---|---|---|
| 24 hours | ₩12,000 | ₩19,600 |
| 3 days | ₩20,500 (crossed-out original price ₩23,000) | From ₩29,000 |
| 1 weeks | ₩28,500 (original price ₩32,000) | ₩27,900–49,000 |
| 2 weeks | ₩44,000 (original price ₩49,000) | — |
| 4 weeks | ₩62,000 (original price ₩69,000) | — |
The store description includes a line: "no need to buy mock-test sets separately." One pass covers unlimited tests for its duration. That line targets Riiid's Santa SAY, a big-company product sold per set: ₩10,000 for 1 sets and ₩29,900 for 5 sets (launch prices, per Yonhap 2023-09).
The name is deliberate too: "토스트" (toast) is "토스" plus one character, so test-takers searching "토스" in the store will easily find it.
Small volume, decent money
Start with Google Play (install counts are third-party estimates):
| Date | Total installs | About |
|---|---|---|
| 8-18 | Launched (same day on iOS) | — |
| 9-10 | 626 | About 50 |
| 10-08 | 2,021 | About ¥50–80 |
| 10-31 | 3,314 | About ¥40–60 |
| 11-10 | 3,790 | About ¥45–55 |
For three months, daily new installs stayed between 24 and 85 with no spike-and-crash, which doesn't look like big paid UA. On that few installs, 11-09 it still charted #38 in the Korea Play education Top Grossing, 11-10 #43 overall. Charting Top Grossing on so few installs means a high share of installers pay, and pay a lot each.
The real money is on iPhone. Korean test-takers skew iPhone, and TOAST's ratings count on the App Store went from 10-28 385 to 453 in November (4.9 stars). Since 10-28, it has sat between 22 and 40 on the Korean education Top Grossing chart every day. 11-10 ranks #24, right next to the same developer's vocab app 「26초」 (#25). The top spot, 토스미, sat at #17–#27 over the same period. Nobody outside can look up how much money those ranks translate into per month.
Who's behind it
The developer, Apeiro (아페로), is based in Gangnam, was founded in 2021, has 9 people, and raised a seed round in December 2023 (thevc.kr). Their previous product is the vocabulary app "26초" (26 seconds). According to Platum 2026-05-20, a startup news outlet, it reached 500k users in its first year, was profitable for 6 consecutive months, and grew 33% a month on average. Its Play version has over 280k installs, and the traditional Chinese version has 5,098 ratings on the Taiwan App Store.
TOAST is the same team running the same playbook again, this time in a category with higher ticket prices. The users are students and job seekers preparing for the TOEIC, the developer account is the same, and the promotion process is the same (see next section). I found no direct evidence that "26초" cross-promotes TOAST.
Where users come from
Sock-puppet reviews on Instagram, confirmed. At least 6 Reels tagged "#협찬" (sponsored) tag the official @tos_trainer and link tostrainer.com, with nearly identical captions: "토스트는 AI 채점으로 진행돼서 오히려 부담 없이 학습할 수 있는 게 장점" (TOAST uses AI grading, so the advantage is you can study without the pressure). It is clearly a script sent out to everyone. In Sep 27, an account called "대학생활 모음집" (College Life Collection) also posted image-and-text content. 26 seconds In January this year, the same trick was used with "26 seconds app discount benefits #협찬."
A neat row of five-star reviews in the first week. From Aug 20 to 27, the App Store showed 10 long five-star reviews, each describing a different scenario: studying for graduation requirements, urgently needing a score for work, one week before the exam, studying at home ("집순이 갓생템"), studying better than books ("책보다 낫다"), and leaving online lectures running without watching. Starting from Aug 28, the reviews change in tone (see below). Whether this first-week batch came from a review-experience program or incentivized reviews can't be determined from outside (an inference, since the wording and timing are suspicious).
Store search. The name borrows from "토스", and the subtitle packs in "토스 모의고사·템플릿·발음 피드백" (TOSS mock test, templates, pronunciation feedback). Searching "토익스피킹 앱 AI 채점 추천" (TOEIC Speaking app AI scoring recommendation), the top two results are its App Store and Google Play pages.
Not found: no press coverage, and no discussion on X or TikTok. Meta's ad library doesn't index its page. Not finding it doesn't mean it wasn't advertised.
One more timing point: it launched in mid-August, just before the exam prep season ahead of the second-half public hiring rounds at Korean companies (9–November).
Who it's competing with
| Competitors | Scale | How it differs from TOAST |
|---|---|---|
| 토스미 TOSME | On iOS since 2024-09, 1,288 ratings 4.8, Korean Education Top Grossing 17–27. Claims 80k learners (2025-09 reported by etnews as 50k). | The category leader. Same playbook but about 30–40% more expensive. Sells on "1-second score results" and question prediction; the web version also has a refund-if-you-hit-target class. |
| Santa SAY(Riiid) | A big company, launched 2023-09. | Sold per set, not unlimited for a period. |
| 토선생 | On iOS since 2026-08-23, 16 ratings. | Launched the same week as TOAST, not gaining traction yet. |
| 토스업, 토스하이, 토스스, etc. | Some are 2020-year-old legacy products, others are small indie apps. | Thin experience or unclear scale. 토스하이 also recruits promoters on Instagram. |
| YBM official past-exam practice, entor | entor is free for 2 attempts, then ₩1,000 per attempt. | Official and pay-per-attempt cheaper alternatives. |
Next door, OPIc (another common English speaking test in Korea) has spawned a dozen-plus "오픽 AI 모의고사" (OPIc AI mock test) apps this year: 오픽미, 오픽마스터, 오픽코치, 오픽이지… (App Store, Korea store). In Korea, both tests' AI mock exam space is already packed.
What the people who paid are complaining about
After the first week's five stars faded, most of the reviews left on the App Store are two or three stars, and the complaints boil down to four things:
- Expensive. Aug 28: "1 free mock tests and then it's all paid… I looked at the price and deleted it".
- Few questions, repetitive formats. Sep 5: "The mock test formats are all the same, and the words repeat too". Sep 13, someone gave an example: in Part 4, the table questions keep coming back to the same "workshop schedule".
- Scoring is too rigid. Oct 10: phrasings commonly taught in online courses get marked wrong, e.g. "Most of all (X) → First of all (O)". After about 5 mock tests, "it just lowered my confidence, so I stopped using it".
- Score loading fails. Sep 14 and 18: two reviews on the same day say the score never comes out.
These are the spots the lead product 토스미 spent over a year polishing. TOAST took off not because its product was better, but because the category had already been validated, it was about 30% cheaper, and the team had a ready-made promotion playbook.
Can you ride it?
Copying an AI mock exam for TOEIC Speaking or OPIc in Korea is not recommended: 토스미, Santa, and TOAST have already established themselves, several indie developers are crowding in at the same time, and you don't have the ready-made user base and promotion experience that 26초 had.
The part worth taking away is the way it's sold:
- Pick the right exam. Three criteria: the registration fee is high (test-takers benchmark against the registration fee, not against other apps' prices); there's a fixed test date (payments cluster in the one to two weeks before the exam); and the section is speaking or writing, which you can't practice alone to a score (reading alone gives no feedback, while AI scoring is useful).
- Sell short passes, not monthly subscriptions. 24 hours, 3 days, 1 weeks, so people cramming before the exam can buy enough in one go. Add a 4-week pass for those preparing early.
- A free diagnostic is a must. Give an estimated score in 10 minutes so people can see how far they are from their target, then prompt the upgrade.
- Get the question bank and scoring right first. TOAST's negative reviews all point here, and this is where followers can do better than it.
On whether it works in other countries, we only did a light check: the notable English apps in the App Store Japan and Taiwan stores are vocabulary builders, TOEIC listening and reading prep, and AI English conversation (mikan, abceed, Speak, スタディサプリ). We didn't find one that specifically scores speaking against a given exam's rubric and sells short pre-exam passes (inference; we only searched for apps with TOEIC, 英検 or 英檢 in the name). Japan's Eiken second-round interview and Taiwan's GEPT speaking section both meet at least two of the three criteria above, but we haven't verified how large the demand is. Apeiro's own 26초 already has a Traditional Chinese version, so moving TOAST to Taiwan and Japan is only a matter of time.
If it were you
B. Worth trying, but not in Korea. The "short pre-exam pass + AI speaking scoring" model has been shown to make money in Korea, and few people outside are paying attention yet. The weak spots: each exam's market is small, scoring has to be accurate, you have to understand the exam itself, and the Korean market is already crowded.
Who it suits: 1–2 people, one of whom has taken or taught a local speaking or writing exam (Japan's Eiken second-round interview, Taiwan's GEPT speaking section, or the interview stage of some professional certification), knows what the scoring rubric looks like and where test-takers lose points; the other can wire speech recognition to a large-model scorer and tune it until test-takers trust it.
Who it doesn't suit: anyone who wants to build TOEIC Speaking or OPIc in Korea; anyone who wants a general AI English conversation app (Speak and Speakbuddy already own that lane); anyone who doesn't know the target exam's rubric and plans to let a large model score freely (TOAST's two complaints, "scoring is too rigid" and "off-topic answers still get points", would land on you).
How to build it: pick one exam, break the official rubric and public past-paper formats into a question bank, build at least 10 mock tests like TOAST's plus templates for each question type, and prepare several scenarios per type (not all "workshop schedules"). Launch on iOS first, offer a free 10-minute diagnostic that outputs an estimated score, then sell four unlimited-access tiers: 24 hours, 3 days, 1 weeks, and 4 weeks, priced against the registration fee. Scoring results should explain why points were deducted, accept common correct phrasings, and not mark down expressions taught in online courses.
How to get users: copy TOAST's three moves. Put the abbreviation test-takers search for in the name and subtitle. In the first few weeks after launch, find university students and job-hunting accounts on Instagram (in Japan, swap in prep accounts on X and YouTube) and sponsor reviews. Build App Store reviews from real users' post-exam feedback. Time ad spend to the exam calendar, with the budget concentrated from registration deadlines through the exam weeks.
When to launch: ship 4–6 weeks before the next sitting of the exam you chose, so you catch a full pre-exam peak. To use Korea's rhythm as an analogy: first-half public recruitment runs 3–April, and 1–February is prep season.
Signs you've misread it: ① two or three similar apps for your chosen exam already sit in the local App Store top 50 of the Education chart, meaning the space is already crowded; ② after a small-budget test over one pre-exam peak, the conversion from free diagnostic to paid doesn't cover acquisition costs; ③ TOAST or 토스미 cut prices or switch to monthly subscriptions, meaning the short pre-exam pass model can't hold up even in Korea; ④ the exam owners themselves (ETS, YBM, the Japan Eiken Foundation) launch free or cheap official AI mock tests.
Ads it's running
Meta has shown us 16 ads; 16 are still running, earliest from Oct 2026 5.
Meta · Video · Running for 47 days · Still running · View original ↗
토익스피킹, 오늘부터 앱 하나로 끝내세요IM부터 AH까지 토스트가 책임질게요. 토익 스피킹 목표 레벨, 토스트로 가볍게 도달해보세요. 1. AI채점 모의고사 : 무료로 10분만에 예상 점수 확인 2. 개인화 학습 플랜 : 노베이스부터 AH까지 맞춤형 솔루션 제공 3. 나만의 약점 분석 : 말하는 즉시 빠르게 약점만 집중할 수 있는 피드백 제공 4. 파트별 템플릿 : 보고, 듣고, 말하기 무한 반복 프로세스
Meta · Running for 36 days · Still running · View original ↗
토익 스피킹, 오늘부터 앱 하나로 끝내세요내 토스 예상점수, 시험보기 전에 미리 확인해봐요. 토스트 앱 가입만 하면 무료 모의고사 받아볼 수 있어요!
These are the longest-running ads in this set. Long runs usually mean the creative is working. Data comes from the official Meta, Google and TikTok ad libraries and updates daily; open the official library to view the original creative.
Android store page ↗ iOS store page ↗
Research notes: competitor table, acquisition evidence, full research report
- Competitors
Name Where Launched Scale How it differs from this one 토스미 TOSME(TeacherK) iOS 2024-09 iOS 1,288 ratings 4.8, Korea Education top grossing 17–27; claims 80k learners (2025-09 hours 50k, etnews) Category leader, same model but pricier (24h ₩19,600, 3 days ₩29,000+), pitching 1-second scoring and question prediction; TOAST is 30% cheaper, adds study plans and template-based shadowing Santa SAY(Riiid 뤼이드) Web 2023-09 Big-tech AI education company; at launch 1 tests ₩10,000, 5 tests ₩29,900 Priced per test; TOAST is unlimited within the period 토스업(HUE6) iOS 2020-07 iOS 57 ratings 3.1 Older product, poor experience, little traction 토스하이 TOSHIGH Android Before/after 2025 (unverified) Unverified; ran Instagram 서포터즈 promotions Similar AI mock tests, scale unknown 토선생 iOS 2026-08-23 16 iOS ratings Indie product launched the same week, no traction YBM TOEIC Speaking 기출 체험 / entor AI 모의고사 Other Early Official agent / web per-attempt ₩1,000 Official real-test practice and cheap per-attempt pricing, no unlimited-within-period option ChatGPT voice / online courses (해커스, YBM) Big-company features — — Usable for practice, but doesn't give estimated scores per the TOEIC Speaking rubric - Where users come from
- Verified Instagram micro-influencer #협찬 reviews: at least 6 #협찬 Reels tagged @tos_trainer and linking tostrainer.com, with near-identical copy (e.g. instagram.com/reel/DffBDr-T1bo, DffAkEZhUQz, Dfe9qb_KZmt); 2026-09-27 more posts from university-life accounts; the earlier 26초 used the same playbook 2026-01
- Inference Launch-week reviews are clustered: in the App Store review sample from 08-20 to 08-27, 10 are long 5-star reviews with uniform wording. After September, real reviews are mostly 2–3 stars (too expensive, scoring too strict, too few questions). The app got 385 ratings within 71 days of launch.
- Inferred Store search / ASO: the name "토스트" echoes "토스", and the subtitle reads "토스 모의고사·템플릿·발음 피드백"; ranks in the top two for searches like "토익스피킹 앱 AI 채점 추천"
- Inferred Cross-traffic from its own 26초 users: same developer account; 26초 has 284k Play installs, 6,954 iOS ratings, and claims 500k users, all TOEIC test-takers; no direct evidence of in-app cross-promotion found
- Inferred Paid ads (Meta / Apple Search Ads): not tracked in the Meta ad library, nothing found; 0 related posts on TikTok and X
- Verified Season / timing: 60% of TOEIC Speaking test-takers are job-hunting; launched 8-18, ahead of Korea's second-half public recruitment (9–November); one-time registration fee ₩84,000
Why TOAST (토스트, TOEIC Speaking AI mock test) worked
- Package
com.apeiro.speakpass, iOS id 6796267457; developer Apeiro (아페로, Seongnam, founded 2021-05, 9 people, 2023-12 took seed funding; thevc.kr) - Channels: Play + iOS launched on the same day (2026-08-18), Korea only, Korean only
- Research date 2026-11-09/10
Verdict
TOAST is Apeiro's prior vocabulary app "26초" playbook run again in a new category: "TOEIC Speaking (토스) pre-exam cram". The category was already validated by 토스미 (TOSME): test-takers buy unlimited AI mock tests by the 24 hour, 3 day, or 1 week, pay in the one to two weeks before the exam, and the average order value is high. TOAST followed by cutting prices 30–40%, adding diagnostics, study plans, and template shadowing, flooded launch week with positive reviews, then used small Instagram accounts to post #협찬 reviews.
It makes its money mainly on iOS: ranked 22–40 on Korea's Education top grossing chart, in the same tier as 토스미 (17–27) and its own 26초. Play has only 3,790 installs at about 50 per day, did not chart before October, and only squeezed into the Korea Education top grossing chart at #38/43 when 11-09 came. This is a small, profitable vertical exam app, not a hit; it got going mainly because the category was validated early and the team had a ready distribution playbook, while the product itself is a modest improvement.
Data
| Metric | Value | Source |
|---|---|---|
| Play cumulative installs | 626(09-10)→ 2,021(10-08)→ 3,314(10-31)→ 3,790(11-10) | googleplay library app_day_ext (AppstoreSpy estimate) / app_day |
| Play daily installs | Stayed between 24–85 with no spike; roughly 43–57/day over the past month | Same as above |
| Play charts | Korea EDUCATION top grossing 38 (11-09), 43 (11-10); the library had no charting record before 10-29 | googleplay library chart_rank |
| iOS ratings | 385 (10-28) → 450 (11-10), +2 to +11 per day; App Store page shows 453 ratings, 4.9 stars | appstore library app_day; App Store kr page |
| iOS charts | Korea Education top grossing 22–40, charting daily since 10-28 (11-10 at #24, right beside its own 26초 at #25); Korea Education top free #71 | appstore library; App Store page |
| In-app purchase prices (iOS kr) | 24 hours ₩12,000; 3 days ₩20,500/23,000; 7 days ₩28,500/32,000; 2 weeks ₩44,000/49,000; 4 weeks ₩62,000/69,000 (each tier shows two prices, likely a discounted price plus the original) | App Store kr page |
| Play in-app purchase range | $8.99–$49.99 per item | googleplay library |
Bottom line: most of the revenue comes from iOS (Korean test-takers skew heavily to iPhone, and iOS has far more ratings and better chart positions than Play). Play reached the top grossing chart with 3.8k installs, which suggests a high conversion rate and high order value, but the absolute revenue is small. What chart rank translates to in dollars was not verified.
Competitors
The TOEIC Speaking (Korean nickname 「토스」) AI mock test category has been crowded since 2023:
| Product | Platform | Launched | Scale | How it differs from TOAST |
|---|---|---|---|---|
| 토스미 TOSME(TeacherK / Hajin Jung) | iOS + Android + web | 2024-09(iOS) | iOS 1,288 ratings 4.8, Korea Education top grossing 17–27; claims "80k learners", 2025-09 hours 8 months 50k people (etnews) | Category leader; TOAST copies its model directly. Same selling approach (unlimited short-term passes), but pricier: 24h ₩19,600, 3 days ₩29,000–39,000, 1 weeks ₩27,900–49,000. Its pitch is "1-second scoring + practice questions", plus a web-only refund program (환급반, pays back if you hit your target). TOAST is cheaper and adds study plans and template-based shadowing. |
| Santa SAY(Riiid 뤼이드) | Web / inside the Santa app | 2023-09 | Big company (Riiid is an AI education firm that raised over ₩100 billion); 1 sets ₩10,000, 5 sets ₩29,900 (at launch), later 5 sets ₩39,900 | Sold per set, not as a time-based monthly plan; TOAST uses "unlimited within the period" on top of per-set pricing |
| 토스업(HUE6) | iOS + Android | 2020-07 | iOS 57 ratings 3.1 | Older product, poor experience, little traction |
| 토스하이 TOSHIGH | iOS + Android | Around 2025 | Runs a "서포터즈" (supporters) Instagram promotion | Also uses micro-influencer promotion; scale unverified |
| 토스스(com.sooyang.toeicspeaking) | Android | Not verified | Unverified (not in our database) | Small product from an individual developer |
| 토선생 | iOS + web | 2026-08-23 | 16 iOS ratings | Launched the same week as TOAST, not gaining traction yet. |
| YBM TOEIC Speaking 기출 체험 / 엔토르 entor AI 모의고사 | App / web | Early | YBM is South Korea's official TOEIC agent; entor is free for 2 uses, then ₩1,000 per use | Official and cheap pay-per-use alternatives |
| Hackers / YBM online courses, ChatGPT voice | Online courses / big-company features | — | — | No dedicated scored mock exam |
Growth tactics (evidence)
- Instagram influencer #협찬 review (confirmed). At least 6 #협찬 Reels turned up, all from @tos_trainer and all linking tostrainer.com. The captions match almost word for word ("토스트는 AI 채점으로 진행돼서 오히려 부담 없이 학습할 수 있는 게 장점"), which means they're a single brief handed out to every creator: instagram.com/reel/DffBDr-T1bo, DffAkEZhUQz, Dfe9qb_KZmt, DffCFRhgG5f, Dfer-HKJRpn, DfesLegJTOJ. There's also an image post from the college-life account 「대학생활 모음집」2026-09-27: instagram.com/p/Dd0QZznj3l3. YouTube Shorts has the same thing from a small alt account, @Fishbbang, reposting @tos_trainer videos with single-digit views. The predecessor app 26초 2026-01 ran the same playbook: a #협찬 post titled 「26초 앱 할인 혜택」 (instagram.com/p/DT6au-rExy0).
- Positive reviews clustered in the launch week (looks like review prompting; unverified). In the App Store review sample, 08-20 through 08-27 has 10 long 5-star reviews with a uniform style ("독학파면 괜찮을 듯", "집순이 갓생템", "책보다 낫다") covering many scenarios (graduation, job hunting, cramming a week before the test, studying at home). Real reviews after September are mostly 2–3 stars: complaints about price ("한 달에 6 만원"), scoring that's too strict, repetitive question types, and scoring that fails to load (appstore library reviews). As of 10-28 it has 385 ratings; over the 71 days since launch that averages about 5.4 per day, with many concentrated in the first few weeks.
- Store search / ASO (inferred). The name 「토스트」 sounds like 「토스 (TOEIC slang)」, and the subtitle packs in "토스 모의고사 · 템플릿 · 발음 피드백" (TOEIC mock tests, templates, pronunciation feedback). When searching "토익스피킹 앱 AI 채점 추천" (TOEIC Speaking app AI scoring recommendation) on Jina, TOAST's App Store and Play pages rank in the top two.
- Cross-promotion to the parent's 26초 users (inferred; no direct evidence found). Both apps are published under the Apeiro Corp. developer account, and their users are students and job seekers preparing for the TOEIC. 26초 has 284k installs on Play and 6,954 ratings on iOS, and claims 500k cumulative users. No evidence of in-app cross-promotion was found.
- Paid ads (unknown). The Meta Ad Library doesn't track Apeiro's page; that doesn't mean they aren't spending. On TikTok there are 0 related posts (TikHub); on X there are 0 (x-radar). No media coverage.
- Seasonality (timing). 2025 of TOEIC Speaking test-takers 60% are doing it for job hunting, and 25–29-year-olds make up 52.1% (Korea TOEIC Committee; newstap / daum reports). A single test registration costs ₩84,000 (namu.wiki). The app launched in mid-August, just before the study rush for the second-half public job postings (9–November).
Why it's working
- The category is proven and people pay for it. 토스미 has spent over a year showing that "short-term unlimited mock exam passes bought before the test" can hold a spot on the Korea iOS education Top Grossing chart for a long time, ranking in the top 30. Students compare it against a one-time exam registration fee of ₩84,000 and tutoring that costs hundreds of thousands of won, so passes priced at ₩12,000–69,000 are an easy sell.
- Fast follower, lower price. The same "unlimited N days" pass: TOAST 24h at ₩12,000 (토스미 ₩19,600), 3 days at ₩20,500 (토스미 ₩29,000). On features, it adds a 10-minute free diagnostic, a daily study plan generated from your target score and exam date, and 200+ question-type templates with shadowing and instant error correction.
- The team has a playbook already. 26초 reached 500k users in its first year, turned profitable for 6 consecutive months, and grew 33% month over month (Platum 2026-05-20). Same 9-person team, same developer account, same cold-start process of Instagram sponsorships plus store reviews, now applied to an exam category with higher price points.
- iOS and Korea only. It hit Korea's education Top Grossing chart in the top 40 on iOS right after launch, with Play as a side channel. Play's daily growth has stayed around 50, with no sign of chart-pushing, suggesting steady organic traffic plus small-scale promotion rather than big paid user acquisition.
Differentiation
TOAST's product differences come down to three things: a diagnostic before a study plan, 200+ question-type templates with shadowing and instant feedback (more about practice than just testing), and unlimited-within-period membership that's 30–40% cheaper than 토스미. Its own marketing says the LLM explains why points were deducted; 토스미 and Santa SAY offer that too.
These differences are not the main reason it took off. Users are still buying "unlimited mock exams before the test + score estimate," same as 토스미. The bad reviews focus on a small question bank, repetitive question types, scoring that's too strict, and loading failures, which suggests the product still trails the leader. It took off because the category was already validated, the price was low, and the team had distribution experience. Price is the only difference that likely made a real impact.
Visibility
- TOAST itself: no media coverage, X and TikTok discussion at 0 levels, revenue not public, only a batch of Instagram sponsored posts. It's still not widely known.
- The parent company 26초's user numbers and profitability were reported by Platum (2026-05), and 토스미's learner count was reported by etnews (2025-09). The category is no secret in Korea's education startup scene.
Copycats
Haven't seen anyone copying TOAST specifically. But a group of individual developers crowded into the same category around the same time: 토선생 (2026-08-23) and 토스스. In the neighboring OPIc category, over 2026 years a dozen-plus "OPIc AI mock exam" apps appeared (오픽미, 오픽마스터, 오픽코치, 오픽이지, and others; appstore library).
What others can borrow
- Find a local vertical exam with "exam + time limit + high registration fee" (TOEIC Speaking, OPIc, JLPT, various professional certifications). Sell short passes like "24 hours / 3 days / 1 weeks unlimited" to people about to take the test, rather than a monthly subscription. Test-takers spend most willingly in the last week or two before the exam.
- Follow a leader whose model is already proven, cut the price by about 30%, add a "diagnostic → plan" flow to make the first experience smooth, and at launch run Instagram micro-influencer sponsorships and store reviews.
- A small team uses one developer account to ship several vertical exam apps in a row, sharing the user base and promotion process. Once one app turns profitable, they copy it to the next exam.
- Similar exams could be replicated in other countries: Japan's TOEIC S&W / Eiken 2nd-stage interview, Taiwan's GEPT oral test, and others (demand not verified).
Still unclear
- How much monthly revenue corresponds to iOS Top Grossing ranks 22–40; how revenue splits between Play and iOS.
- Whether it's spending on Meta, Apple Search Ads, or Naver ads, and how big the budget is (not tracked in ad libraries).
- Whether 26초 sends users to TOAST inside the app.
- Whether the 5 star reviews in the first few weeks after launch came from a tester group or incentivized reviews (the writing style and timing are suspicious).
- Retention and repeat purchases: negative reviews mention a small question bank, so it's unclear whether users will buy again before their next test.
Sources
- Play page: play.google.com/store/apps/details?id=com.apeiro.speakpass
- App Store kr: apps.apple.com/kr/app/id6796267457 (rating 453, price table, reviews)
- Website: tostrainer.com/
- 26초: platum.kr/archives/287210, thevc.kr/apeiro, skimmify.com/
- 토스미: www.tosme.kr/home, apps.apple.com/kr/app/id6502824958, www.etnews.com/20250910000039
- Santa SAY: www.yna.co.kr/view/AKR20230926159400017, kr.aitutorsanta.com/santa-say
- Instagram sponsored posts: see links above; official account www.instagram.com/tos_trainer/
- TOEIC Speaking test-taker profile: www.newstap.co.kr/news/articleView.html?idxno=321480, namu.wiki/w/TOEIC%20Speaking%20and%20Writing%20Tests
- Our database: googleplay (chart_rank, app_day, app_day_ext), appstore (app_day, review, v_chart), as of 2026-11-10; metaads has no records; TikHub and x-radar have no related content.