
₩200k Surges to "Grossing #2": How an AI Usage Widget Hacked the Chart, and What Happened After
What it is It shows how much quota is left on subscriptions like Claude Code, Codex, Grok and Gemini, and when it resets, as widgets on iPhone, a Mac menu bar item and Apple Watch. Monthly ₩3,900, yearly ₩19,000, one-time ₩39,000.
Why it took off In the Korean developer-tools grossing chart a few subscriptions are enough to reach the top. It spent about $140 on Meta ads to push onto the chart, took Korean search terms like "클로드 한도" (Claude limit) that nobody was targeting, and a first-launch paywall lifted conversion about 3×. After charting, the creator 조코딩's main channel featured it as a case study.
CHow you could follow it Don't build yet another AI usage tracker. Instead, take its four steps (pick a low-bar category chart, push with small paid spend concentrated in a few days, long-tail ASO in your native language, and a first-launch paywall with a 7-day trial) and apply them to another small tool people check several times a day.
Est. monthly revenue (App Store) under $1KWe estimated this from its chart ranks and rating growth across storefronts, calibrated against verified revenue; typical error is 2–3x. Read it as a tier, not an exact number.
- Data (updated daily) · launched 2026-08-21 · US store rating 0 · charted in 11-20 of 5 stores
iOS · best category chart rank · by store US storeKoreaJapanData 2026-10-28 → 2026-11-20 (24 days, 24 observations). Lower rank is higher; gaps mean it wasn't charted that dayiOS · rating count · US store Data 2026-11-16 → 2026-11-20 (5 days, 3 observations). Apple doesn't publish download counts; rating growth is a proxy for downloads.
조코딩, a Korean YouTuber who coding tutorials, had a "product builder" at his company build a small tool that shows how much quota is left on subscriptions like Claude Code and Codex, and when it resets, as widgets on iPhone, a Mac menu bar item and Apple Watch. Three weeks after launch it reached #2 on the Korean iOS developer-tools grossing chart. Three moves got it there: ₩200k in Meta ads, Korean search terms nobody else was bidding on, and a paywall that asks you to choose a subscription the moment you open it. The product itself is about the same as dozens of rivals, and its rank has already dropped from #9 to #69. The playbook is worth learning; the product isn't worth following.
Oct 6 update
On Oct 6, 2026, 조코딩 (real name Dong Keun Jo, 752k YouTube subscribers) posted across LinkedIn, Threads, Instagram and the YouTube community, titled "조코딩 빌더, 앱스토어 '매출 2위' 찍은 3가지 전략" (Three strategies a 조코딩 builder used to hit #2 in the App Store grossing chart).
The post lays all its cards on the table: launched 3 weeks ago, the iOS developer tool hit #2 Top Grossing and #5 Top Free in the Korean store, held the top spot for 9 days straight, and 10; it reached #5 on the Korean Mac chart, and charted in 11 countries including the US and Japan. How did they pull it off?
- Meta ads spent about ₩200K (around $140) to get 275 target users, at ₩729 per install. "It was the spark that got us charted, and organic traffic took off after that";
- grab keywords like "클로드 한도" (Claude quota) and "AI 사용량" (AI usage) that people search for but few apps target, and rank #1 in search;
- A month later: 4.6, it ranks #12.8 on the Korean developer-tools grossing chart, with 1.8 ratings worldwide, and hasn't gained in the last 11.3 days.
Check back in a month: 11-09 is #69 on the Korea Developer Tools Top Grossing chart, with 6 ratings worldwide and no new ratings in the last 7 days.
What it is
The problem it solves is specific. Claude Code and Codex subscriptions have two limits: one every 21 hours and one weekly. Heavy users run several accounts at once and have to time their "reset tickets" carefully. The app pulls remaining quota and reset times for Claude Code, Codex, Grok, Antigravity and Gemini into one panel, warns you when you're close to running out, and lets you use reset tickets straight from your watch. Credentials sit in the user's own iCloud Keychain and never touch their servers.
The problem it solves is very specific. Claude Code and Codex subscriptions have two layers of limits: one every 5 hours, and one weekly. Heavy users often run several accounts at once and have to time their reset tokens carefully. This app pulls remaining quota and reset times from Claude Code, Codex, Grok, Antigravity, and Gemini into one dashboard, alerts you when you're running low, and lets you use reset tokens right from your watch. Credentials stay in the user's own iCloud Keychain and never touch the company's servers.
| Price (Korean store) | |
|---|---|
| Free | Only supports 1 accounts. A 9-25 one-star review says: "Only 1 can be linked, the rest require a paid upgrade" (only one, the rest have to be paid). |
| Monthly | ₩3,900 |
| Annual | ₩19,000 with a 7-day free trial |
| One-time purchase | ₩39,000 |
Updates are frequent: 8-21 through 11-07 shipped about 25 versions, adding a watch app, a Mac menu bar, iCloud sync and a service status page. 10-23 added a Report page in 1.1.0 that converts your token usage into money at API prices, and can generate a "receipt" you can share. The latest 1.1.2 follows Claude's plan renames (Max 5x is now Max, Max 20x is now Ultra).
The good reviews are all about multiple accounts and resets: "리셋 가능한게 개꿀이네요" (being able to reset is great, 9-18); another says "Many apps don't support multiple accounts... the standby mode while charging works too, I keep it next to my computer as a usage monitor" (10-12).
How big is "Top Grossing #2" really?
Very small. This chart's bar is remarkably low.
- On the same Korean developer-tools Top Grossing chart, Limits: AI Usage Tracker has only 2 ratings in Korea, sitting at #15 in 10-28; TokenMaxxer has only 3 ratings and ranks #4 in 11-07. So "#2 Top Grossing" probably only takes a few subscriptions a day (my inference, no revenue figures are public).
- Its trajectory on the Korean Top Grossing chart (App Store developer tools): 10-28 #50,10-29 #9,10-30 #15,10-31 #22,11-01 #27,11-03 #40,11-05 #50,11-07 #60,11-09 #69, steadily falling. The timing fits: the push up came as people on the 7-day trial converted, and it slid once the trial wave passed (inference).
- Ratings: 6 worldwide (5 in Korea, averaging 4.0; 1 in Argentina). The store page says it was "recently reset", so real reviews may be a bit higher, but not by an order of magnitude.
- Android: 100+ downloads on Google Play (11-08).
- The only new signal: 11-09 it entered the Japanese developer tools Top Grossing chart for the first time, at #20, with 0 ratings in the Japan store.
It has never charted on the overall chart in any country.
Where users come from: three moves, plus one brand
Small paid push to chart. ₩200K bought the first batch of users to lift the ranking, then the chart's organic traffic took over. A Facebook page called "FlowWorks" 10-13 posted a 28-second video titled "조코딩 팀이 만든 AI 사용량 트래커" (AI usage tracker made by the 조코딩 team), 4.6K views, and it looks like the creative used for this batch of ads. The ad creative uses 조코딩's name directly.
Grabbing Korean keywords nobody else targets. English-speaking developers wouldn't build for terms like "클로드 한도" (Claude limit), but Koreans search exactly those words. The store name and subtitle are stuffed with keywords.
A 7-day trial on the first-launch paywall. Conversion nearly 3x'd, and on a chart where a few sales are enough, that alone was enough to reach #2. The version history matches: 1.0.11 (9-18) added a 7-day trial to the annual plan, and 1.0.17 (10-01) added an onboarding screen. Also, 1.0.18 (10-02) asks for a rating once you've connected your account.
Jocoding is the brand behind these reports. They post the same report across four platforms, and the creator account 「터닝클럽」10-27 uses their short videos as the flagship case for "one person shipping 3–4 apps a month." But note the order: the report went out after the app was already charting, so the big accounts amplified traffic that was already there. The real starting point was paid acquisition and keywords.
It's not a one-person effort either. JoCoding is hiring "Product Builders" with a base salary of ₩40M plus revenue share, and the same developer account has shipped about 20 apps since 2026 March; this is one that landed.
Almost nobody in the English-speaking world knows it: it doesn't show up on X, Reddit or Hacker News.
Who it's competing with
The category is already crowded. About 100 AI usage tracker apps have been listed since 2025-10, and more than 40 of them came after 2026-08.
| Competitors | What it is | Compared with alternatives |
|---|---|---|
| CodexBar | Open-source Mac menu bar tool, from 2025-11 | GitHub 23k stars; the site says it supports 106 providers and is free; no iPhone widget or watch app |
| Limits: AI Usage Tracker | iOS, listed 2026-07-14 | Charted in 65 countries, 46 ratings in the US store; nearly identical features, the widest reach |
| TokenMaxxer: AI 토큰 사용량, 한도 확인 | iOS, Korean indie developer, listed 10-12 | Puts "한도 확인" (limit check) right in the name, with unlimited accounts and 46 languages; 11-09 ranks #11 on the Korean developer tools Top Grossing chart, ahead of it |
| UsageScope, UsageMaster | iOS, Korean developer, listed 2026-03 and 2026-06 | Did Korean first, but on the Top Free chart, never pushing for Top Grossing |
| Official | Claude's Settings › Usage page, and Claude Code's /usage command | Shows a single account, not multiple accounts or providers combined; no widget |
After the win announcement, a cluster of similar apps appeared in the Korean store: Limi (10-06), TokenMaxxer (10-12), Kitty AI Usage (10-16), VibeMon (10-19), Token Tracker (10-31), Kettle (11-02). Names and keywords look alike, but it's unclear whether they were built after seeing the announcement.
No room in other languages either. Across the developer tools charts in 11-09 countries, local-language clones already hold positions: "nowdex - AI 用量追踪" is #22 in the China store, "Usages:AI 用量追蹤" is #34 in the Taiwan store, "UsageNow: KI-Nutzung Tracker" is #36 in the Germany store, and "Nomo - AI Status & Usage" is #21 in the Italy store. In the Japan store, the highest-ranking one is this app itself, at #20.
What you can learn is the playbook, not the product
Its features are covered by CodexBar, Limits and TokenMaxxer, each of which matches more than half of them. What got it onto the chart wasn't the features but four things: choosing a chart where a few sales are enough and that you can talk about afterward; buying the first users for $140; taking Korean terms English-speaking developers never see; and using the paywall to concentrate trials into the first few days so they convert.
The playbook could travel to other categories, provided their Top Grossing charts have a similarly low bar. The catch is that 조코딩 has already explained it publicly, everyone in Korea's solo-founder scene has seen it, and "Top Grossing #2" probably amounts to a small income plus a good story. For JoCoding that story has value: it helps recruiting and sells courses. For you, the story may not be worth much.
A cap hangs over the whole category too: if Anthropic or OpenAI someday adds a usage-limit widget and multi-account summary to its official app, dozens of these apps get squeezed together.
If it were you
C: You can learn a playbook for climbing charts, but this product isn't worth it. The category already has about 100 similar apps, localized into Korean, Chinese, Traditional Chinese, German and Italian, and it just took the top spot in Japan itself; even its own ranking is slipping. It got this far on distribution and 조코딩's reputation, and you can't borrow that brand.
Who it suits: people who want to practice the "small budget, climb one country's category chart" playbook at 1–2 scale, who have a small tool for developers or heavy users that gets checked several times a day, and who can write the store page and keywords in a language English-speaking developers won't bother with.
Who it doesn't suit: anyone wanting to build yet another AI usage tracker, since this space is packed; anyone hoping to make real money from the category, since "Top Grossing #2" likely means a few sales a day; and anyone with no audience of their own who won't spend a cent on acquisition.
How to do it: take its four steps and apply them to another category. First find a low-bar category Top Grossing chart: if several apps in the top 20 have single-digit ratings, the bar is low. Then find a number people check several times a day and make it a widget, menu bar item or watch app; these tools are naturally high-frequency. Put the paywall on first launch with a 7-day trial, and feature the annual plan most prominently.
How to get users: first spend a few hundred thousand won (a few hundred dollars) on Meta ads to concentrate installs in the first few days and push into the chart; at the same time, pack the store name, subtitle and keywords with native-language long-tail terms. Once you're on the chart, screenshot your rank and use it as creative, posting it on Threads, LinkedIn and local developer communities.
Is the window still open?: For AI usage tracking, the window has already closed. The playbook itself isn't time-sensitive, but once the official apps ship a limit widget, any tool built on other companies' subscription limits will lose value at once. If you can use a different category, do that instead.
What would show this read is wrong: (1) its Korean ranking climbs back into the top 10, or its Japan ranking rises from #20 and ratings start growing, suggesting the category has more repeat purchases and localization room than it looks; (2) JoCoding publishes revenue at several million won a month or more; (3) one of the copycats listed after October holds a top-10 spot on a country's developer tools Top Grossing chart for 30 consecutive days while its ratings climb into the dozens, showing latecomers can still break in and stay.
Research notes: competitor table, acquisition evidence, full research report
- Competitors
Name Where Launched Scale How it differs from this one CodexBar(steipete) Open source 2025-11 GitHub 23,061 stars; the site says it supports 106 providers Free and open source with many more providers, but only a Mac menu bar app; it fills the gaps of the iPhone widget, watch app and Korean language Limits: AI Usage Tracker(getlimits.app) iOS 2026-07-14 Charted in 65 stores; 46 ratings in the US store (appstore library) Nearly identical features and broader reach; its strengths are Korean localization, multiple accounts, reset coupons and sync across three platforms Usage for Claude + ChatGPT iOS 2025-11-19 Not verified Earlier, but only supports Claude / ChatGPT; this one covers Grok, Antigravity, Codex and more UsageScope: AI 사용량 관리 iOS 2026-03-06 8 ratings in the Korean store, ranking #41–93 on the Top Free chart Did Korean earlier, but on the Top Free chart and never pushed for Top Grossing; this one used a hard paywall and paid acquisition to climb the Top Grossing chart TokenMaxxer: AI 토큰 사용량, 한도 확인 iOS 2026-10-12 11-07 #4 on the Korean developer tools Top Grossing chart A latecomer, focused on unlimited accounts / API keys and 46 languages; same Korean keywords, and it now ranks ahead of this app UsageMaster - AI 사용량 관리 iOS 2026-06-03 #44–197 on the Korean Top Free chart, 0 ratings Same set of Korean keywords, but without the paid acquisition and paywall playbook AI 사용량 - 클로드·코덱스 한도(app.nlap.aiusage) Android Not verified Not verified Same Korean keywords, Play Store only Claude Settings › Usage / Claude Code /usage Big-company features 2025 All Claude users The official apps show usage for a single account, but not multiple accounts or providers combined, and have no widgets or alerts - Where users come from
- Verified Meta small-budget ads: official 10-06: roughly 200k won reached 275 target users, at 729 won per install. The FB page FlowWorks posted the video "AI usage tracker built by the jocoding team" with 10-13K views 4.6.
- Verified Store search / Korean ASO: officially claims to have reached search rank 1 for "클로드 한도" and "AI 사용량"; the Korean store name is "AI 사용량 트래커"; current rankings not rechecked (Apple search API 403)
- Verified Hard paywall on first launch + 7-day trial: officially: subscription (including trial) conversion went from 4.6% to 12.8%, and trial start rate from 1.8% to 11.3%; the annual plan got a 7-day trial in 1.0.11, and an onboarding screen was added in 1.0.17
- Verified Prompted reviews: 1.0.18 (10-02) shows the App Store rating request after the account is connected
- Verified founder accounts / media: 조코딩 has 752K YouTube subscribers and 20.6K Threads followers, and posted 10-06 results across four platforms (1.1K views on Threads); 터닝클럽 10-27 short videos. The posts went up after it had already charted, so they amplified a chart position rather than creating it.
- Inferred Promo codes for the Lifetime plan: 1.0.7 supports App Store promo codes redeemable for Lifetime Pro; the purpose is unverified
AI 사용량 트래커 / AI Limits Tracker(JoCoding, Inc.,iOS 6801493876)
Regions scanned: App Store library, full-region sweep (115 stores). The app charted in kr / jp / ar / br / eg / pl / sv / tw, eight in total. Focus on kr (home market) and jp. Similar apps' distribution across all stores was also scanned.
One line
Korea's top programming YouTuber 조코딩 (752k subscribers) has a 'Product Builder' from his company build this tool: it puts the 5-hour / weekly usage windows, reset times and 'reset tickets' for Claude Code, Codex, Grok, Antigravity and Gemini subscriptions onto an iPhone widget, Mac menu bar and Apple Watch. Monthly ₩3,900, annual ₩19,000, one-time purchase ₩39,000.
Verdict: it ranked on distribution, not product. The Korean iOS 'Developer Tools' Top Grossing chart has a very low bar (apps with 2–3 ratings could make the top 20). It pushed its way in with three moves: about ₩200k in Meta ads, ASO on Korean long-tail terms like 'Claude limit / AI usage', and a hard paywall on first launch. Once charted, 조코딩 himself promoted it across several platforms as a recruiting and course case study. Its features are nearly identical to open-source CodexBar, Limits on iOS and dozens of other clones. The absolute revenue behind the chart rank '매출 2위' is probably very small. Its rank has slid one way since late October, and newer clones have already passed it.
1. Data: how far it actually got
- Official update (Jokoding 2026-10-06, posted to Threads / LinkedIn / Instagram / YouTube community at once): at 3 weeks since launch, the app was #2 Top Grossing and #5 Top Free in the Korea iOS developer tools category, charted for 9 straight days, 10 days ago. #5 on Korea Mac. Charted in 11 countries including the US and Japan. Threads, LinkedIn
- What our library shows (appstore library, app_day, kr, Developer Tools Top Grossing): 10-28 #50 → 10-29 #9 → 10-30 #15 → 10-31 #22 → 11-01 #27 → 11-03 #40 → 11-05 #50 → 11-07 #60 → 11-09 #69, a steady slide. It first charted on Japan's Developer Tools Top Grossing at #20 (jp rating 0), 11-09; Top Grossing in Argentina 119; Brazil, Egypt, Poland, Sweden and Taiwan only hovered around #100–200 on the Top Free chart. It has never charted on any overall chart.
- Ratings: 6 worldwide (kr 5, ar 1), growing 0 over the last 7 days; kr average 4.0. The App Store page shows the rating was 'recently reset' at some point, so the real review count may be a bit higher, but the order of magnitude doesn't change.
- Google Play: 100+ downloads (Play page, read 11-08); it's not in our googleplay library (it never charted there).
- Pricing: Monthly Pro ₩3,900, Annual Pro ₩19,000, Lifetime Pro ₩39,000 (kr store page). The free tier only lets you connect 1 accounts (a one-star review on 9-25: 「1개만 연동되고 나머지는 유료결제 필수」; a review on 10-12 says the free tier allows 2, so the limit may have been adjusted).
- How low the chart bar is: on the same chart, Limits: AI Usage Tracker has only 2 ratings in kr and ranks #15 with 10-28; TokenMaxxer has only 3 ratings in kr and ranks #4 with 11-07 (appstore library). So 'Korean Developer Tools Top Grossing #2' may take only a few subscriptions a day (inference, amounts can't be verified).
2. Competitors: who else is doing the same thing
This is an extremely crowded category. In the full appstore library catalog (about 1.8M apps), filtering by name for 'AI / Claude / Codex / Token + usage / limit / quota / 사용량 / 한도 / 使用量' turns up about 100 clones released after 2025-10, and more than 40 of them released after 2026-08. Direct competitors:
| Product | Where | Launched | Scale | How it differs |
|---|---|---|---|---|
| CodexBar(steipete) | Open-source macOS menu bar apps | 2025-11 | GitHub 23,061 stars (gh api, 11-08); the site claims support for 106 providers | Free and open source, with far more providers; Mac only, no iPhone widget or watch |
| Limits: AI Usage Tracker(Pranav K. M.,getlimits.app) | iOS | 2026-07-14 | Charted in 65 stores; 46 ratings in the US store (appstore library) | The widest reach and most ratings on iOS; nearly identical features, no Korean localization |
| Usage for Claude + ChatGPT(amir hayek) | iOS / Mac / Watch | 2025-11-19 | Not verified | Earlier; covers only Claude and ChatGPT |
| CUStats Go | iOS | 2026-01-22 | Charted in 27 stores, #1 in a Netherlands category chart (appstore library) | Small, single-digit ratings |
| UsageScope: AI 사용량 관리(Jaemin Song) | iOS (Korean developer) | 2026-03-06 | 8 ratings in kr, Top Free #41–93 (appstore library, November) | Earlier with Korean-language targeting; went the Top Free route and never broke into Top Grossing |
| UsageMaster - AI 사용량 관리(MyeongJin Lee) | iOS (Korea) | 2026-06-03 | #44–197 on the Korean Top Free chart, 0 ratings | The same Korean keyword set |
| TokenMaxxer: AI 토큰 사용량, 한도 확인(Minjun Kim) | iOS (Korea) | 2026-10-12 | 11-07 is #4 on kr's Developer Tools Top Grossing, already ahead of it (appstore library) | Pitches 'unlimited accounts & API keys', 46 languages; puts 'check limit' right in the name |
| AI 사용량 - 클로드·코덱스 한도(app.nlap.aiusage) | Android | Not verified | Not verified | Same Korean keywords, Play Store only |
| Chrome extensions like Claude Usage Monitor | Chrome | From ¥2025 | several | Browser-only |
| Official feature | big tech | From ¥2025 | — | Claude has a Settings › Usage page and Claude Code has a /usage command; you can see your usage, but there's no multi-account roll-up, no widget and no alerts |
Verdict: no empty market. When it entered (8-21), CodexBar and dozens of iOS clones already existed globally, and UsageScope and UsageMaster were already in the Korean store.
3. Growth tactics: where users come from
- Small Meta ad cold start (confirmed): the official line is that it spent about ₩200k, got 275 target users at ₩729 per install, and this was 'the trigger for climbing the charts, after which organic traffic took off' (Threads / LinkedIn 10-06). The Facebook public page 'FlowWorks' ran a 28-second video 10-13 titled '조코딩 팀이 만든 AI 사용량 트래커' with 4.6K views (Facebook). It looks like creative from this campaign, and it borrows 조코딩's name. Our metaads library doesn't track this advertiser (the library only covers a watchlist, so not finding it doesn't mean it didn't run ads).
- Korean long-tail ASO (confirmed as the official claim; current rank not verified): it targeted terms like 'Claude limit' and 'AI usage' that have search volume but few competitors. The official claim is that it reached search rank #1. The Korean store name is 'AI 사용량 트래커', with the subtitle '메뉴바에서 보는 AI 한도'. Apple's search API returned 403 today, so the current search rank couldn't be re-checked.
- Hard paywall on first launch (confirmed): after switching from 'use free first' to 'pick a free trial or subscription as soon as you open the app', subscription conversion (including trials) for new installs rose from 4.6% to 12.8%, and trial starts rose from 1.8% to 11.3% (LinkedIn 10-06). The version history matches: 1.0.11 (9-18) also gave annual plans a 7-day free trial; 1.0.17 (10-01) added an onboarding screen. Trials converting in the first 10 days likely explains the early-October climb to #2 Top Grossing; the steady slide since then matches the trial wave fading (inference).
- Review prompt (confirm): 1.0.18 (10-02), "After you link your account, we'll ask you to leave a rating on the App Store."
- Jocoding own traffic & media (confirmed): YouTube 752k subscribers, Threads 20.6k followers, the win report posted on four platforms at once (the Threads post got 1.1K views with 10-06). The short video from the founder-interview channel 「터닝클럽」10-27 used it as the showcase for "a company that hires one startup person as Product Builder and ships 3–4 apps a month" (Instagram). There is also an Android Shorts clip ("AI tool made by the Jocoding team"). But the win report went out after it had already charted, so this traffic is an amplifier after the chart, not the starting point.
- Lifetime redeem codes (possible): 1.0.7 (9-9) supports redeeming an App Store offer code for Lifetime Pro inside the app. It may be used to hand out codes in communities; the purpose is unverified.
- X / Reddit / HN: in the x-radar corpus over 90 days, 'AI Limits Tracker' and 'jocoding' each appear in 0 posts; nothing turned up on Reddit or HN. It has essentially no presence in English-speaking communities.
4. Why it took off
- The demand is real, and it's intense, Korean-style demand: since 2025, Claude Code and Codex have had two layers of limits, '5 hours + weekly', and heavy users run several Max accounts at once and have to time their 'reset tickets'. The good reviews land on exactly these points: 「리셋 가능한게 개꿀」「다중계정 등록 안되는 앱들 많았는데… 충전 중 스탠바이까지 지원해서 사용량 모니터처럼 컴퓨터 옆에 띄워두고 씁니다」 (appstore library, kr reviews 9-17 and 10-12).
- It picked a chart with a very low bar: Korea's iOS Developer Tools Top Grossing only takes a handful of subscriptions to reach the top 10, and '#2 Top Grossing' is a number you can put in a pitch.
- Acquisition cost was pushed very low: about ₩200k in ads plus Korean ASO, at ₩729 per install; the hard paywall lifted conversion by about 3×.
- It has backing: JoCoding's Product Builder program pays a base salary of ₩40M plus revenue share (hiring page, AI Times). The same developer account has published about 20 apps since 2026-03 (appstore library), and this app was one successful shot among them. After it charted, 조코딩's main accounts promoted it as a case study, which drew another wave of follows.
- Iteration pace: from launch on 8-21 to 11-07 it shipped about 25 versions, adding Watch, Mac menu bar, iCloud sync, a status page, and a Report page (converts token usage to API prices and lets you share a 'receipt'). This supports retention, but it isn't why users arrive.
5. Differentiation
Compared with Limits and CodexBar, what it adds is: multiple accounts, niche providers like Grok and Antigravity, reset tickets usable directly in the app and on the watch, iPhone + Mac + Watch synced via iCloud, Korean localization, plus API cost reports and 'receipt sharing' from 1.1.0. But none of these are unique on their own: CodexBar supports more providers and is free, Limits' widget and countdown are nearly the same, and TokenMaxxer leads with unlimited accounts. So differentiation isn't what drove it. The drivers were Korean keywords, a low-bar chart, the hard paywall, and the 조코딩 brand. The feature with the most viral potential is the 'receipt sharing' on the Report page, but it only launched 10-23, and I see no sign yet that it's driving growth (ratings grew 7 over 0 days, and rank kept sliding).
6. How many people see it now, and whether anyone is copying it
- Visibility: it's been thoroughly covered in Korean developer and solo-founder circles. 조코딩 published its ranks, cost per install, conversion rates and keyword approach, and 「터닝클럽」also ran a feature. But it hasn't disclosed revenue figures, and it's barely mentioned on English X, Reddit or HN.
- Copycats: after the 10-06 win announcement, a string of similar apps from Korean developers appeared in kr: Limi - AI Usage Widget (SION NOH, 10-06), TokenMaxxer (Minjun Kim, 10-12, puts 'check limit' right in the name, #4 kr Developer Tools Top Grossing 11-07), Kitty AI Usage (chul kim, 10-16), VibeMon: 코딩 에이전트 모니터 (Jungyoul Yu, 10-19,11-09, #30 kr Developer Tools Top Grossing), Token Tracker: AI Usage (Taewoon Kang, 10-31), Kettle: AI Usage Reset Alerts (Hyun Jun Lee, 11-02). The timing and keywords look like copying, but there's no evidence it was inspired by this announcement.
- Current status: Korean Top Grossing rank 9 → 69 (10-29 → 11-09), ratings stuck at 6; Japan just charted at #20 in Developer Tools Top Grossing (11-09), a small new signal.
7. Takeaways
- Look for demand that comes from numbers people check several times a day, and turn it into a widget, watch app or menu bar item. That kind of product is high-frequency by nature.
- The cheap bet that makes a good story: buy the first batch of installs in some small country's top-grossing category for a few hundred thousand won, then use the chart rank as marketing material.
- Use the local language to grab the long-tail keywords English-speaking developers don't cover.
- A hard paywall at first launch with a 7-day trial can lift conversion 3× on niche tools.
- The studio model: "solo founder × base salary + revenue share × high-frequency testing": one account ships 20 apps in half a year, and when one hits, they pitch it.
8. Still unclear
- Actual subscriber counts and revenue (Apple doesn't disclose them; the barrier is low, so presumably very small).
- Current search ranks for "Claude limit" and "AI usage" in the Korean store (Apple search API 403, not re-verified).
- Whether Meta ads are still running and in which countries (not tracked by the metaads library).
- Rankings for the Mac menu bar version (the database only has the iPhone charts).
- Japan 11-09 charting: unclear whether they started buying Japan traffic or it's organic.
- If Anthropic or OpenAI builds multi-account usage tracking into their official apps, this whole category gets squeezed.