Stringers
Mobile App · iOS · written 11-18
프시케 - 사주, 만세력, 타로, 운세를 풀어주는 AI · SOOYEON KIM

One question, four fortune-telling systems side by side. Psyche, an AI fortune app by a Korean indie dev, hit the Top Grossing chart within a week, on Meta ads and a ₩66,000 credit pack

What it is An AI fortune chat app: the same worry gets read side by side through Four Pillars, tarot, Western astrology, and Zi Wei Dou Shu, with sources cited and follow-up questions allowed. You buy conversation credits per use, ₩4,900 to ₩66,000 per pack.

Why it took off It started running Meta install ads the day after launch, and it had set up paid-conversion tracking through a web quiz before the app even shipped. The Korean Lifestyle category's Top Grossing bar is low, so a few big purchases on a high-priced credit pack are enough to push it up the ranks.

CHow you could approach it Don't copy this product in Korea. Copy the playbook for Japanese fortune-telling or Traditional Chinese Zi Wei and Bazi markets: prove the tracking and acquisition cost with a web quiz first, then launch the app. Free for one perspective, paid for multi-perspective comparison, and consumable credit packs.

Grade C: the playbook is worth learning, the product isn't worth copying. Korean AI fortune apps are already crowded, and Horii launched a multi-system comparison half a year earlier. It has only 11 ratings averaging 2.9, its rank comes from paid acquisition and big purchases, and it hasn't proven it can hold. Few people know about it yet

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) · First launched 2026-11-12 · Korean rating 25 · 11-20 charted in 1 stores
iOS · best category chart rank · by store
2030405011-1711-1811-1911-2052
Data 2026-11-17 → 2026-11-20 (4 days, 4 observations). Lower rank is higher; gaps mean it wasn't charted that day
iOS · rating count · Korea
010203011-1711-1811-1911-2025
Data 2026-11-17 → 2026-11-20 (4 days, 4 observations). Apple doesn't publish download counts; rating growth is a proxy for downloads.

A Korean indie developer launched the AI fortune app Psyche (프시케) in Nov 12. The same worry gets four readings side by side: Four Pillars, tarot, Western astrology, and Zi Wei Dou Shu, and you can keep asking follow-ups. A week later it reached #23 Free and #32 Top Grossing in the Korean App Store's Lifestyle category. It didn't take off because of the product, though. Side-by-side fortune readings were already around, and most first-week reviews were complaining about errors. What's worth studying is the ad and monetization setup it built a month before launch.

"Retried 30 many times"

In Nov 17, the day after the app updated to 1.0.3, the Korean App Store brought in 5 new 1–2 star reviews. One reads: "I rarely write reviews, but this time I actually tapped 'retry' 30 times. The AI doesn't understand what I say, it cuts off halfway through the answer, and keeps telling me to refresh. Worst." Another: "From login to filling in my details, nothing saves, and asking a question never produces an answer, even after retrying." One is just a single line: "Nothing works properly."

The same day, it first appeared on the Korean Lifestyle chart: #69 Free, #52 Top Grossing. On day two Free climbed to #23, on day three Top Grossing rose to #32, and the rating count jumped from 1 to 11, averaging 2.9 (Korean App Store chart and ratings, 11-17 to 11-19).

A flood of one-star reviews on one side, rising ranks on the other. The explanation lies in its ads and pricing.

What it is

First you fill in your birth details, then talk about your situation the way you would in a chat: should I change jobs, what does my boyfriend really think? You can pick a single perspective, or have it compare up to 3 perspectives side by side on the same question, then list where they agree and where they differ. Every answer cites its basis: which pillar in the chart, which card, which aspect in the natal chart. Terms you don't understand can be followed up on. It remembers what you've told it before, and for compatibility readings it can put two people's profiles side by side. The model underneath is Google's Gemini, as shown in the privacy policy and the web code.

The developer's name on the App Store is SOOYEON KIM. Google Play lists a sole proprietorship called 모락 in Yeonsu-gu, Incheon, and its mail-order sales registration number was only issued in 2026. This developer account previously published two other apps, one for learning English and one for practicing TOEFL speaking, each with only 1 ratings.

Where the money comes from: conversation credits, largest pack ₩66,000

No subscription. You buy consumable "conversation credits" that never expire:

PackPricingAbout
5 times₩4,900$3.4
15 times₩12,900$9
45 times₩29,900$21
120 times₩66,000$45
Comparison questions 5 times₩5,900$4

(App Store in-app purchases and official pricing, 11-18; USD converted at about ₩1,450)

There's a small daily free allowance, but free use only lets you pick one perspective. New users get a batch of paid credits at sign-up, and the first answer in onboarding doesn't use up a credit. Side-by-side comparison only works with paid credits: the free tier lets you taste one reading, and if you want to see "what the other systems say" you have to pay. The paywall sells "more angles, more confidence" rather than "unlock the full version," which suits fortune-telling users well.

The bar for Korea's Lifestyle Top Grossing chart is low. 11-19 ranked #30 to #38 are 사주월드, 음양관, 사주핑, 우주고양이 보라, and 마이파이, mostly small fortune apps with tens to hundreds of ratings. A single ₩66,000 pack is worth as much as a dozen small purchases from others, so a handful of buyers on the big pack can push the rank up. Judging from the rank, it's probably making a few hundred dollars a month (an estimate, with a wide range). "Top Grossing #32" sounds impressive, but it's roughly at that scale.

Where users come from: ads started the day after launch

The main channel is Meta app install ads. The landing page is myomyo.ai, a newly opened page with 0 likes. 11-19 is running 6 ads, all with a "Install Now" button:

  • Since Nov 13, 1 videos linking to Google Play, using a user-voice script: "A Four Pillars reading that actually predicts my boyfriend's family situation? Honestly more useful than just a compatibility score... I really need to try it ㅋㅋㅋ"
  • Since Nov 18, 5 image ads linking to the App Store and Google Play, with headlines "내 만세력을 쉬운 말로" (my Four Pillars chart in plain words), "Psyche, chatting with Four Pillars and tarot", and "What do tarot cards mean? Ask Psyche."
  • Placements covered Facebook, Instagram, Threads, WhatsApp, and Audience Network.

(Meta Ad Library, 11-19)

iOS ads started on 11-18, and iOS rankings climbed from 11-17 to 11-19 over those same days, so the timing lines up.

What's more notable is that it wasn't spending blind. About a month before the app launched, the same team was already running a psychology quiz on the web: TCI temperament and personality, ADHD, impostor syndrome, and Four Pillars love, each returning a free report with paid detailed versions at ₩9,900 / ₩19,900. The site's terms of service first appeared in Oct 23, and its domain certificate dates to Oct 6. The privacy policy says events like "onboarding completed," "first conversation completed," and "credit purchase (with amount)" are sent back to Meta. The web code also contains an internal dashboard that calculates the "acquisition cost per free report" daily, broken down by traffic source across Instagram, Threads, Naver, Kakao, and TikTok. In other words, once the app went live, ads could target people directly based on who had paid.

The rest are small channels: invite codes giving both sides 4 credits, daily check-in credits, and credits for writing feedback. A recruit-a-superfan program offers a pack worth about ₩30k. Inside the app there's a "Cheer on Psyche" entry point that prompts users to leave a store review. How many of the 10 ratings received on 11-19 came from that entry point can't be determined.

It also runs a Threads content account, @psyche_park00, posting daily relationship-psychology jokes paired with Four Pillars, Zi Wei, and astrology. Its bio reads "Try it once and you'll see. Way better than GPT and Gemini, right?" 11-19 It has only 51 followers, and each post gets 0 to 2 likes, bringing in almost nobody. No coverage was found in media, on X, or on TikTok.

Who it's competing with: Korean AI fortune apps are already a red ocean

About 13 Four Pillars and tarot apps charted in the Korea Lifestyle Top Grossing top 40 on 11-19. Closest to 프시케:

CompetitorsLaunchedScale (App Store, Korea storefront, 11-19)Compared with Psyche
호리2026-05325 ratings, Lifestyle Top Grossing #8Multi-system consultation across Four Pillars, tarot, astrology, and Zi Wei. Launched half a year earlier and ranks far higher.
김귀자 할머니2026-049,782 ratings, 7 day +156, Top Grossing #17A "fortune grandma" persona chat that relies on the character and word of mouth.
사주마녀2025-1230 ratings, grossing #6Also a low-rating-count app that climbed on paid acquisition and big purchases.
아임미리2026-04699 ratingsAI Four Pillars and tarot plus psychological counseling, very close positioning.
점신, 포스텔러, 헬로우봇2015–201715k to 125k ratingsAn established player selling report packages, live human consultations, or scripted chat.

There's also a free competitor: ChatGPT and Gemini. Threads and Korean forums are full of posts about pasting Four Pillars, Zi Wei, and birth charts into GPT for analysis. One app called Waitcle does just one thing: it generates prompts for you to paste into ChatGPT. Psyche's ad lines, "Four Pillars in plain words" and "understands fortune better than GPT," target people who already use GPT to read their charts.

So "multi-system comparison plus follow-up questions" isn't its moat. Horii did it earlier and bigger, and a Chinese developer's Orrery launched in June with the same combination and still has zero ratings.

What you can learn is the playbook, not the product

  • Build the web quiz with its conversion tracking and acquisition-cost dashboard first, then launch the app. On day one of the app, you can optimize ads against paid events, saving the tuition of spending blind.
  • Give one perspective for free, charge for comparison. The paywall sits on "a few more takes," which helps hesitant users most.
  • Consumable credit packs, with the top tier priced high. Fortune users tend to buy large packs on impulse. On a low-bar category chart, a few big purchases can buy visibility.
  • Ride the habit users already have. People are already pasting their charts into GPT, so the ads just need to say "I'll lay out your chart, explain it in plain words, and show you the basis."

The downside is worth remembering too: ads pull people in, but if the back end can't keep up, one-star reviews follow, and an average of 2.9 drags down organic traffic and conversion afterward.

If it were you

C. You can learn a complete "web quiz → conversion tracking → paid acquisition → credit packs" playbook, but copying this product itself isn't worth it. Korean AI fortune apps are already crowded, the features it sells were done by Horii first, and it hasn't proven it can hold. Two or three days of rank plus 11 ratings averaging 2.9 can't establish demand.

  • Who it suits: a small team of 1–2 people who can run Meta install ads and do the acquisition-cost math, working in a market with strong fortune demand where AI fortune apps aren't yet crowded (such as Japanese "占い", Traditional Chinese Zi Wei and Bazi; Southeast Asia hasn't been checked). The team should understand fortune-telling a little and be able to write readings and ad copy that don't read like machine translation.
  • Who it doesn't suit: anyone aiming for Korea (13 similar apps on the same chart); anyone who can't run paid ads and is counting on store search or content-account organic traffic (Psyche's Threads account with 51 followers is the example); and anyone whose back end can't handle concurrency, since users brought in by ads who hit an error will leave a one-star review.
  • How to do it: don't build the app first. Build a free personality or chart quiz on the web, offer a free short report, sell the detailed version in the ₩1–20k range, and wire up the Meta pixel with server-side conversion events. Run it until you know the cost per free report and per payment. Once the numbers make sense, build the app and keep the "one perspective free, multi-perspective comparison paid, consumable credit packs" model. Before launch, stress-test the big problems: truncated long answers and retry loops.
  • How to acquire users: Mainly Meta install and conversion ads. Copy two creative styles: user-voice short-form videos that feel like organic recommendations ("Turns out it can match your boyfriend's family situation"), and explainer images that turn your chart into plain language. Landing page and App first screen both say: "Easier than copying into GPT yourself, and it shows its reasoning."
  • When to do it: Before the window closes. There's no platform window and no rush, but general-purpose LLMs are getting better at fortune-telling, so the "we do the chart for you" value is shrinking. Early is better than late. If web tests run two months and paid-acquisition cost per payment still isn't below the average order value, don't add money to the App.
  • What would show I'm wrong: If by 2027 January, 프시케 is still holding a top-30 spot in Korea's Lifestyle Top Grossing, ratings climb to several hundred with an average back above 4, that means "multi-system comparison" can carry paid conversion even in a red ocean, and it's worth upgrading and copying its playbook into other countries. Conversely, if it drops off the chart within a week of stopping Meta spend, its ranking was just bought with ads.

iOS store page ↗

Research notes: competitor table, acquisition evidence, full research report
Competitors
NameWhereLaunchedScaleHow it differs from this one
호리(Yumeta Lab)iOS2026-05-29iOS 325 ratings, kr Lifestyle Top Grossing #8 (11-19); Play estimated ~$13K/monthSame kind of multi-system consultation (Four Pillars, tarot, astrology, etc.), and older and bigger. 프시케's angle is side-by-side comparison, sourced answers, and follow-up questions. The difference is small.
김귀자 할머니 - 사주 캐릭터챗iOS2026-04-19iOS 9,782 ratings, +7 in 156 days, Lifestyle Top Grossing #17; Play estimated ~$36K/monthCharacter chat built on a "fortune-telling grandma" persona. 프시케 uses a neutral, counseling-style tone plus multiple systems.
점신iOS2015-02iOS 113,994 ratings, estimated iOS revenue ~$62K/monthVeteran all-in-one fortune platform with live human consultations. 프시케 is AI chat only.
포스텔러iOS2017-07iOS 14,983 ratings; Play estimated ~$118K/monthLarge fortune content library, buy reports one by one. No multi-system comparison or follow-up questions.
헬로우봇iOS2017-12125,583 ratings, estimated iOS revenue ~$42K/monthMulti-character, script-style chatbot fortune-telling. Not free-form follow-up questions.
사주마녀iOS2025-12-2330 ratings, but Lifestyle Top Grossing #6 and Top Free #17Four Pillars, synastry, tarot consultations. Similar: few ratings, relies on paid acquisition and big purchases to climb the charts.
아임미리iOS2026-04-17699 ratingsAI Four Pillars, tarot, and psychological coaching. Positioning is very close, but no side-by-side comparison.
술술(Aves)iOS2024-09iOS 1,461 ratings, Meta running 21 adsSend a letter and wait for a reply, marketed as "free". 프시케 is instant chat with paid multi-perspective answers.
Waitcle / OrreryiOS2025-09 / 2026-061,347 / 0 ratingsWaitcle only generates prompts for users to paste into GPT. Orrery offers the same multi-system mix but never took off.
ChatGPT / Gemini plus "copy AI reading" tutorialsBig-company features—Threads, Daum, and X are full of tutorials on "copy your Four Pillars, Zi Wei, and natal chart into GPT together".Free substitute. 프시케 does the chart and multi-system pairing for the user. The ad pitch is simply "understands fortune-telling better than GPT".
Where users come from
  • Verified Meta app install ads: Advertiser page myomyo.ai (page_id 1359134740623975,0 likes) has 6 live ads: starting 11-13, 1 user-voice videos ("남친 집안 사정까지 맞추는 사주가 있다고?…") landing on Play; starting 11-18, 5 images ("내 만세력을 쉬운 말로" etc.) landing on App Store and Play; placements FB/IG/AN/WhatsApp/Threads (Meta Ad Library, pulled via ScrapeCreators 11-19)
  • Verified Conversion postback / CPA optimization: Release notes for version 1.0.3 (11-16) say "integrated ad measurement". The privacy policy says onboarding completion, first chat, and purchases (with amount) are sent back to Meta. The website JS bundle contains an internal Meta performance and CPA dashboard.
  • Verified Invites, check-ins, feedback, and superfan rewards: Invite codes give both sides 4 coupons each (up to 4 people). Daily check-ins and usage feedback also grant coupons. Choosing "열혈팬" (superfan) gets a coupon pack worth about ₩30K. Version 1.0.3 added an invite link that auto-fills the code (website i18n copy, release notes).
  • Inferred Prompting users to leave store reviews: The App has "프시케를 응원해 주세요 → App Store에서 리뷰 남기기" (Support 프시케 → leave a review on the App Store), saying it will be "rewarded with more perks". Rating +11-19, 10. Not verified whether coupons were traded for good reviews.
  • Verified Threads content account @psyche_park00: Posts daily dating-plus-Four-Pillars, Zi Wei, and astrology jokes. Bio link carries utm_campaign=history_stories. 11-19 has only 51 followers, 0–2 likes per post. Minimal impact.
  • Inferred Store keyword search: App name is stuffed with 사주, 만세력, 타로, 운세, AI. Effect can't be quantified.

프시케 (Psyche) AI fortune chat (iOS 6830268906) research report

In one line: an AI fortune-chat App built by a Korean indie developer (SOOYEON KIM, sole proprietor name "모락"). A single worry can be read from four angles (사주/만세력 Four Pillars, tarot, Western astrology, Zi Wei Dou Shu). The answers from each angle can be compared side by side, and you can keep asking follow-ups. Payment is per-use consumable "chat coupons", ₩4,900–66,000. 2026-11-12 on iOS (Android released simultaneously, package name app.psyche.ai). Since 11-17 it has charted in Korea's Lifestyle category, both Top Free and Top Grossing. What pushed it up the category charts was Meta app install ads: running on Android from 11-13, on iOS from 11-18. The Threads content account mentioned in the recall card has only 51 followers and is not the main driver. Add to that Korea's low bar for the Lifestyle Top Grossing chart, where a few people buying big packs can push the rank up. The product itself isn't new: similar "multi-system AI fortune-telling" apps already exist, including 호리 and several others. Reviews in the first week mostly complain about errors and lag, with a rating of 2.9. Volume is still small, so calling it "proven" is premature.

Data dates: appstore library 2026-11-17 to 11-19; Meta Ad Library (via ScrapeCreators), website, store pages, and Threads were captured on 2026-11-18/19.

1. How to read the data card (correcting first)

  • Ranks (appstore.chart_rank; only the kr store, all in the Lifestyle category 6012, never charted overall):
    • 11-17: Top Free #69, Top Grossing #52
    • 11-18: Top Free #23, Top Grossing #43
    • 11-19: Top Free #92, Top Grossing #32
    • The recall card says "Top Free #26", which doesn't match the library. The library's best is #23 in 11-18.
  • Rating count: 11-17 and 11-18 both at 1 ratings, 11-19 jumped to 11 (+10), average 2.9 (appstore.app_day; iTunes lookup kr, 11-19). 10 new ratings in one day with the average stuck at 2.9 means a lot of 1-star ratings. The reviews match: 11-17 got 5 reviews in one day at 1–2 stars, all saying "keeps asking me to retry," "AI replies cut off halfway," "login and saving profile both error," "slow after the update." Only 11-13 and 11-17 each have one 5-star review (appstore.review).
  • What a Top Grossing #32 app is worth: 11-19 The Korean Lifestyle Top Grossing chart's #30–38 spots are 사주월드, 음양관, 사주핑 and 우주고양이 보라 and 마이파이. Most of them are small fortune-telling apps with a few dozen to a few hundred ratings (App Store library). Stringers' App library estimates the one at this rank makes about $326 a month (range $125–7,206), so treat it as a ballpark only. The recall card says "revenue is climbing fast", but that's overstated: at the top consumable in-app purchase tier of ₩66,000, a handful of big purchases is enough to push a category Top Grossing ranking up to #30.
  • It's new, but what it does isn't: On the same day, the top 40 of Korea's Lifestyle Top Grossing included about 13 Four Pillars/tarot apps: 사주마녀 #6, 김귀자 할머니 #17, 플립톡 #19, 사주GPT #23, 타이트사주 #27, 자모 #28, 사주월드 #30, 음양관 #31, 프시케 #32, 사주핑 #35, 우주고양이 보라 #36, 만세력 8자어때 #37, 마이파이 #38. 호리, ranked #8 in Lifestyle Top Grossing, is also in the same category (appstore library kr 11-19).

2. What it is

  • Developer: App Store seller name SOOYEON KIM (김수연). The Google Play listing gives a sole proprietorship "모락" in Yeonsu-gu, Incheon, business registration no. 465-21-02189, mail-order sales registration 2026-인천연수구-1867 (registered in 2026). Contact email [email protected], and the same email name maps to the site moneytoy.kr (an ASMR points-for-gift-voucher page called "왁뿌볼"; not verified whether the same person built both). Two earlier Apps from this developer account, 인풋영어 (2026-04) and 닥터스픽 (2026-06, TOEFL speaking), each have only 1 ratings (iTunes lookup), so neither took off.
  • Product (store description + copy in the website JS bundle):
    • Create a birth-data profile first, then ask using one or more of the four angles: Four Pillars, tarot, astrology, or Zi Wei Dou Shu. For example, "Should I switch jobs?" or "What does my boyfriend really think?"
    • Multi-angle comparison: For the same question, it automatically picks up to 3 angles and reads them side by side, then lists where they agree and where they disagree.
    • Shows its reasoning: Answers mark which cell of the chart, which tarot card, or which star configuration was used. Terms you don't understand can be followed up on.
    • Memory: Remembers things you've shared before, manageable in settings. For synastry, you can select two people's profiles to read together.
    • Supports voice and attachments (images, PDFs). Models run on Google Gemini (the privacy policy says Google LLC handles "AI chat, analysis, and report generation"; the model enum in the analytics is only gemini_flash / gemini_pro).
    • Targets ages 19 and up; iOS age rating 17+.
  • Pricing (website JS bundle Referrals-*.js, i18n copy):
    • Consumable chat coupons in 4 tiers: 5 uses for ₩4,900, 15 uses for ₩12,900, 45 uses for ₩29,900, 120 uses for ₩66,000 (about $3.4–340). No auto-renewal, no expiry. There is also "comparison questions 5 uses for ₩5,900".
    • Free tier: a daily basic chat quota (only one angle can be chosen); a batch of paid coupons is granted at sign-up; the first answer in onboarding doesn't count. Multi-angle comparison requires paid coupons. This is the paywall hook.
  • There's also a web track: psycheai.app's "나만의방" (My Room) offers psychological tests, including the TCI temperament test (free report about 10 pages), ADHD, impostor syndrome, Four Pillars for love, and life direction. Tests give a free report; the detailed report costs ₩9,900 / ₩19,900, paid on the web, and can be paid through a virtual account. The earliest version of the terms is 2026-10-23, the refund policy version is refund-2026-10-22, and the earliest certificate for the domain on crt.sh is 2026-10-06. In other words, web tests plus Meta conversion tracking were already running about a month before the App launched.

3. Competitors

Korea's fortune-telling App market is crowded and mature, with AI Four Pillars Apps launching one after another over 2025–2026 years:

ProductPlatformLaunchedScale (appstore library kr 11-19 / stringers App library estimate)How it differs from 프시케
점신iOS/Android2015iOS 113,994 ratings; estimated iOS revenue ~$62K/monthAll-in-one fortune platform with live consultations; 프시케 is AI chat only
포스텔러iOS/Android2017iOS 14,983 ratings; Play estimated ~$118K/monthContent-style fortune library, one-off report purchases; no multi-system comparison or follow-ups
헬로우봇iOS/Android2017125,583 ratings; estimated iOS ~$42KMulti-character chatbot fortune-telling, script-style, not free-form follow-ups
김귀자 할머니iOS/Android2026-04iOS 9,782 ratings, +7 in 156 days; Lifestyle Top Grossing #17; Play estimated ~$36KCharacter chat with a "fortune-telling grandma" persona, relying on persona and word of mouth; 프시케 uses a neutral, counseling-style tone
호리(Yumeta Lab)iOS/Android2026-05iOS 325 ratings; Lifestyle Top Grossing #8; Play estimated ~$13KClosest match: four pillars, tarot, astrology, Zi Wei Dou Shu and other systems in one consult. It launched six months earlier and ranks much higher.
사주마녀iOS2025-1230 ratings, yet #6 on the Lifestyle Top Grossing chart and #17 on Top FreeFour pillars, New Year fortune, compatibility readings, tarot consults. Same playbook: few ratings, paid acquisition and big-ticket purchases to climb the charts.
아임미리iOS2026-04699 ratings; its best Lifestyle Top Free rank is #24AI four pillars plus tarot and psychology coaching. Its positioning, "self-understanding," is very close to 프시케's.
술술(Aves)iOS/Android2024-09iOS: 1,461 ratings. Meta is running 21 ads.Letter-writing plus waiting for a reply, pitched as "really free."
WaitcleiOS2025-091,347 ratingsGenerates four pillars, astrology and tarot prompts for users to paste into ChatGPT. 프시케 does that step for them.
OrreryiOS2026-060 ratings, paidAlso covers four pillars, Manseryeok, tarot and Zhouyi across multiple systems. Built by a Chinese developer, but it never took off.
ChatGPT / Gemini plus "copy AI reading" tutorialsBig-company features—On Threads, Daum Cafe and X, there are plenty of posts pasting four pillars, Zi Wei and natal charts together into GPT for analysis.Free substitute. 프시케's Threads bio says, in its own words: "Try it and you'll see it's way better than ChatGPT or Gemini."

Conclusion: "multi-system comparison plus AI follow-up questions" is not 프시케's invention. 호리 was doing it six months earlier and at much larger scale. Waitcle and the many "paste it into GPT" posts show people already meet this need with general-purpose LLMs.

4. How users found it

  1. Meta app install ads (confirmed, the main channel): ScrapeCreators pulled from the Meta Ad Library. The advertiser page is myomyo.ai (page_id 1359134740623975, category Digital Creator, 0 likes, a newly created page). 11-19 has 6 live ads, all with "Install now":
    • From 2026-11-13: 1 videos, landing on Google Play. The copy is user-voiced seeding: "A reading that even matches your boyfriend's family situation? Honestly, something like this is what I actually needed more than a simple compatibility score… seriously, go try it lol".
    • From 2026-11-18: 5 images, landing on the App Store and Play. Headlines: "Explain my Manseryeok in plain words," "Chat with 프시케 using four pillars and tarot," "Ask 프시케 what this tarot card means."
    • Placements: Facebook, Instagram, Audience Network, WhatsApp, Threads.
    • Timing lines up: iOS first charted on 11-17, climbed to #11-18 on the Top Free chart, and ratings rose by +23,11-19 (10).
  2. The tracking setup points to deliberate paid acquisition (confirmed):
    • Version 1.0.3 (11-16) release notes say: "Usage analytics and ad measurement connected per privacy settings."
    • The privacy policy says it uses Meta to measure: landing page visits, onboarding completion, first chat completion, chat ticket purchases (with amount), and free report generation start.
    • The website JS bundle contains an internal dashboard that tracks Meta-reported performance by day, computes "CPA for free report views," and splits traffic by utm and fbclid source (Instagram, Threads, Naver, Kakao, TikTok and others).
  3. Referral and retention incentives (confirmed to exist, effect unknown):
    • 8-digit invite code: both sides get 4 chat tickets, and the inviter can earn up to 4 people's rewards. Version 1.0.3 added "open the app from an invite link and the code auto-fills."
    • Daily check-in gives tickets; submitting feedback gives one ticket.
    • "Superfan" recruitment: selected users get a Standard-tier chat ticket pack, worth about ₩30k.
    • The app has a "Support 프시케" entry that prompts users to leave reviews on the App Store / Google Play. The copy says it will "reward you with more perks," but whether tickets are directly traded for good reviews is unverified.
  4. Threads content account @psyche_park00 (confirmed to exist, minimal impact):
    • Bio: "🔮 Saju, Zi Wei Dou Shu, horoscope reads | Try it and you'll see it's way better than ChatGPT or Gemini?" The link carries utm_campaign=history_stories.
    • Posts a few lines a day on romance psychology plus jokes about four pillars, Zi Wei and astrology, and trivia posts on astrology history (the astrolabe).
    • 11-19 has only 51 followers, with 0–2 likes per post (Threads profile, scraped via 11-19). The recall card treats it as the main traffic channel, but the data can't support that.
  5. Store search (inferred): The app name packs in the keywords "사주, 만세력, 타로, 운세, AI." Many people in Korea search these terms, but competition for them is extremely fierce.
  6. Media, X, TikTok: Nothing found. An x-radar search for psycheai returned 0. A Jina search only turned up appstoretracker listings and Threads posts. Tranco has no ranking (traffic library).

5. Why it made the charts

  • Paid acquisition: Meta install ads started the day after launch, Android first and iOS 5 days later. User-voiced videos plus "plain-language Manseryeok" images target the pain point of "I can't read my chart and want someone to explain it" (Meta Ad Library).
  • Low chart bar: In the Korean Lifestyle category, the Top Free chart ranks #20 or better and Top Grossing ranks #30 or better. An app with two-digit ratings can reach those spots, and they're packed with small fortune-telling apps (appstore library, kr 11-19).
  • High-priced consumable IAP: The top tier costs ₩66,000/120 per session, and multi-system comparison is only available through paid tickets. Fortune-telling users already tend to buy big packs in one go, so a few paying users can push a category grossing rank up (pricing from the website JS bundle; the link between rank and revenue is inferred).
  • The funnel was ready-made: Before the app launched, the web-test funnel already had Meta CAPI feedback, PostHog and a CPA dashboard set up. Once the app went live, it could optimize ads on events like "first chat completed" and "purchase" rather than buying blind (privacy policy, JS bundle).
  • The downside: The backend was unstable in the first week. 1-star reviews focused on errors and retries, with an average of 2.9 (appstore.review 11-17). Users brought in by ads didn't stick, which hurts later organic traffic and conversion.

6. Differentiation

  • Compared with the veterans (점신, 포스텔러, 헬로우봇): it doesn't sell reports one by one. It's a conversation: ask one thing, get an answer, then follow up. Four systems are laid side by side with the basis for each stated. The tone is counseling-style and non-absolute, closer to "using fate-reading to understand yourself."
  • Compared with the new wave of AI four-pillars apps: 김귀자 할머니 and 사주핑 rely on character personas, and 술술 on the ritual of writing letters. 프시케's selling point is "cross-referencing multiple systems and explaining the basis." But 호리 is already doing multi-system synthesis and ranks far higher, and Orrery tried a similar combination without taking off. So the differentiation itself isn't new and isn't the main reason it charted.
  • What actually got it short-term chart placement was distribution and monetization design: Meta install ads, paid tickets for multi-system comparison, large ticket packs, and invite and check-in rewards. The product difference only gave the ads a pitch ("explain Manseryeok in plain words," "understands fate better than GPT").

7. Visibility and copycats

  • Visibility: still low. No public revenue figures, no media or X coverage. The Threads account has 51 followers, Tranco has no ranking, and only automatic listing sites like appstoretracker index it.
  • Copycats: none that specifically follow it. The category was already crowded, so there's no real "who copies whom."

8. What others can learn

  • Using "plain-language chart reading plus the basis GPT can't give" as the ad pitch: Users are already copying Manseryeok into GPT themselves, which shows the need has been educated. The app's job is to package the chart calculation, multi-system comparison and follow-up questions, then push them with a user-voiced seeding video.
  • Build the Meta conversion feedback and CPA dashboard with a web test first, then launch the app: once the app is live, ads can optimize on "first chat" and "purchase" events.
  • Free single-system readings, paid multi-system comparison: charge for "more angles, more confidence," not for "unlocking the full version."
  • Consumable ticket packs, with a very high top tier (₩66,000): on a low-bar category grossing chart, a few big purchases can buy visibility.
  • But don't copy it in Korea: the AI fortune-telling market there is already a bloodbath. Consider moving to Chinese Zi Wei and BaZi markets, or Japan's 占い (uranai) market, with a "multi-system comparison plus follow-up questions" angle.

9. Uncertain

  • How much was spent on ads and whether it paid back (the ad library doesn't show spend); whether the myomyo.ai page is running ads in other countries.
  • Whether the chart placement came from real paid conversions via ads, or from the developer or friends buying large packs to push the rank (no evidence; can only be listed as a possibility).
  • Of the 11-19 extra ratings (+10), how many came from the "Support" review prompt, and whether tickets were traded for good reviews.
  • Android install counts (the googleplay library doesn't index app.psyche.ai, and Jina couldn't pull install numbers from the Play page).
  • Whether the backend stability can be fixed and whether the rank holds; there has been no new version since 1.0.3.
  • Whether the web-test line (TCI, ADHD and others) was ever advertised separately and how much revenue it brought in; not checked.
  • What the relationship is between the moneytoy.kr product and this developer.

Sources

  • App Store:apps.apple.com/kr/app/id6830268906 ;iTunes lookup(country=kr,11-19)
  • Google Play: play.google.com/store/apps/details?id=app.psyche.ai (developer contact info)
  • Website: www.psycheai.app/, /about/terms, /about/privacy (version 11-10), and /assets/Referrals-*.js, /assets/i18n-*.js, /assets/mobile-products-*.js (pricing, tickets, invites, tracking)
  • Meta Ad Library (scraped by ScrapeCreators, page_id 1359134740623975; ads 1451249833005891, 1672110241364753, 2674943069615057, 1093900387018848, etc.)
  • Threads:www.threads.com/@psyche_park00
  • appstore library (agent-im, chart_rank / app_day / review / app_profile, kr, 11-17 to 11-19); Stringers App library (revenue estimates); x-radar; traffic library

This article was written by AI from public data and the data we collect daily. Numbers are sourced, and inferences are labeled.