Launched 25 days before the exam, selling for NT$390: an architecture student's exam-review app for classmates

Stars City Regulations Office is an iOS review app for the "building regulations" section of Taiwan's architect licensing exam, sold as a one-time purchase for NT$390. The developer is an individual dev with an architecture background who promoted it himself on Dcard's architecture board and Threads. Within a week of launch it pushed into the top three of Taiwan's paid education chart. It isn't selling much. It climbed because of the exam season, and because that chart was thin to begin with. The approach could be copied for another exam, but the scale is bound to be small.

A Dcard post

In 2026 Oct 21, someone posted on the Dcard architecture board:

Hi everyone, I made a regulations-reading game app called Stars City Regulations Office. I broke down recent years' regulations questions and picked out the question types that come up again and again and are most worth studying. My goal is simple: consistently pass the 60-point line.

A week earlier, the same person had posted an almost identical post on the same board, for an app called Structure Grand Hotel, which corresponds to the "structure" section of the architect exam. Someone commented under that post: "The regulations exam app is coming soon too~". So the second app had people waiting for it before it even launched.

Both apps are credited to Shangyun Huang and share one account, @arch.app.taiwan, on Threads and Instagram, with the same slogan: "Built something today? The goal isn't a perfect score, it's consistently passing the 60-point line."

What it is

Turns past years' regulation exam questions into a level-based game. A question bank of 1,277 questions, split into tiers and difficulty levels, so a few questions in spare moments counts as a daily check-in. The store page's first line reads "Build a little every day, and 60 passes steadily". The 1.0.4 update shipped in Nov 5 fixed repeated questions: "A new round only starts after all 1277 questions in the bank have come up once" (App Store update notes).

No subscription and no in-app purchases. A one-time NT$390 (about US$12), iOS only.

How much it grew

Developer 10-21 posted on Threads: "Just launched, already #1 in Education." 10-28 Their top spot on the daily charts after that can't be verified; the day they hit #2,10-21 is the best we can confirm.

Don't get fooled by "#2". In the Taiwan paid Education chart, the top 15 are 11-06 apps, and #1 is AnkiMobile, with 131 ratings. #8, #10 and #11 have zero ratings between them. 星城法規局 has only 2 ratings and still ranks #2. It's #97 on the overall paid chart, selling roughly single digits to low teens a day (estimate, not calibrated). At 10 units a day, that's about NT$3,900 in gross revenue per day, and less after Apple's cut. What this chart really shows is that paid Education in Taiwan is empty, not that these apps sell a lot.

Why this moment

The 115 professional architect licensing exam was held Nov 14 to 16 (per the Ministry of Examination announcement). The app launched 25 days before the exam, and the developer's Threads copy kept saying "the last month", "the last week before the exam", "here with you through the final week".

The candidate group for this exam is distinctive:

• Hard: 113 people applied in 4,560, 198 passed, a pass rate of 7.39% (Ministry of Examination statistics, as relayed by Threads users); the 111 pass rate was 5.48%.
• Pass-by-subject, with subject credits kept: many candidates have sat for several years and are only one or two subjects short. Regulations and structure are two of the big subjects that require heavy memorization.
• Willing to pay: people who have paid the registration fee and may have taken prep courses see NT$390 as the price of a meal and barely hesitate.

"Steadily pass the 60-point line, no need for a perfect score" speaks directly to this group: the rule is 60 points per subject to pass, and they're already exhausted from testing. Competitors compete on question count, full coverage, or notes. This one only promises passing.

Where users come from

What can be confirmed comes from the developer's own posts:

1. Dcard architecture board: 10-14 posted in the structure section and 10-21 in the regulations section, both in the "I made one myself" voice, like sharing among classmates rather than advertising.
2. Threads / Instagram @arch.app.taiwan: 10-12 teased "coming soon" on IG. After that, a steady stream of posts like "#1 again" and "#1 in Education" appeared, with single posts reaching about 1,300 views (the 10-19 one).
3. Limited code giveaway: a "nine true, one false" guessing game on Threads, where the first 5 correct guessers received codes by DM. The developer said "about half an hour after posting, the five spots were gone."
4. The two apps cross-promote: structure-app users were the first users of the regulations app. On the App Store "customers also bought" section the two apps appear together, and the competitor Architect Exam Prep (建築師考試通) lists Structure Grand Hotel under "you may also like".

No paid ads, creator deals, or press coverage found, and nobody on X has mentioned "建築師考試" in the past 60 days.

There's a loop worth noting: the Top Paid chart is empty, so a few sales get it into the top five; once it's in the top five, exam takers searching "建築師" in the store see it; then the developer posts the chart screenshot back on Threads, and people in the community see it and buy (inference; no direct evidence for the chart-exposure step). The one-time purchase pricing is what gets it onto the Top Paid chart instead of the crowded Top Free and Top Grossing charts, which is easy to miss.

Users are engaging too. 11-04 one 5-star review asks: "In the normal levels, what's the difference between the levels below and above? Is the difficulty higher?" "Below the question there's a Tier 1/2/3, can you pick the difficulty?" 10-21 reviews want: "Can you add a display showing how many questions have been completed in this round, like the construction one, for a better sense of goal." The developer shipped 4 versions in two weeks and is following the feedback.

Who it's competing with

The one to watch is 《建築師之塔》: same-day launch, same gamification, and self-promotion by people in the architecture community on Dcard. It's not copied; people in the same exam season arrived at the same idea. It suggests "architecture students using AI to write code for classmates" is starting to cluster, and at least for the architect exam, the niche is no longer empty.

If it were you

B, worth a try, but don't expect it to get big. The approach is clear, the cost is very low, and few people know about it. The weaknesses are also clear: the audience for one exam is only a few thousand people, the exam season comes once a year, and one-time purchase revenue stops when the exam ends.

Who it suits: people who have taken or are taking a particular professional exam and have standing in that community; people who can code with AI, are willing to spend a month or two building a small tool, and are happy to earn tens of thousands of NT$ (or tens of thousands of RMB) plus some reputation.

Who it doesn't suit: people who want steady income or want to build a company; people who aren't familiar with that exam and can't post in exam-taker forums as "one of us." Another architect exam app isn't worth it either, since three apps are already competing for the same few thousand people.

How to do it:

1. Pick a different exam. Look for professional exams with low pass rates, pass-by-subject rules, several thousand to tens of thousands of candidates a year, and fixed exam seasons: in Taiwan, technician, CPA, pharmacist, lawyer, and various licensing exams; in mainland China, registered architect, first- and second-level constructor, CPA, and the bar exam follow the same structure.
2. One app per subject, one-time purchase, not a subscription. The goal is to reach the Top Paid chart, where the bar is much lower, and the same person may buy several subjects.
3. The question bank only covers question types that repeat across exams; the copy only promises "passing the cutoff."
4. Launch 1–2 months before the exam, open accounts at the time of the first release to build users, and warm up interest for the next subject.

How to acquire users: post in the forums and communities where exam takers gather (Dcard, PTT, Threads, Xiaohongshu, exam groups), in a "I made this myself" voice; give a few redemption codes to spark discussion; post the chart screenshot back to the community once you rank. No paid ads, since this pool can't support ad spend.

When to launch: ship at least one month before the exam you're targeting; if you miss it, wait until next year. For architects, after Nov 16 this year there's nothing left to do.

What would show you misread it: dropping off the chart after the exam is expected and doesn't count. Misread on the upside: if it stays in the Top 10 Paid Education chart after Nov 16, retake candidates are buying all year, the business isn't just seasonal, and it's worth moving up to A. Misread on the downside: if next exam season sees a dozen similar apps appear at once, or if the same approach can't crack the Top 20 Paid chart for other exams, then this worked only because the architect exam community is small and the chart happened to be empty, which other exams can't replicate, and it should be downgraded to C.