Stringers
Mobile App · Android · written 11-11
Michoo: AI Chat & Roleplay · Dog Games

One line, "🍒 Make your AI girlfriend do anything", translated into 17 languages: a one-person app factory in Vietnam bought 60k installs in two weeks

What it is A generic AI girlfriend / roleplay chat app, Android only. It shows a paywall after a few messages and an ad on every tap, and its store rating is 17+ with sexual content.

Why it took off The developer translated the line "🍒 Make your AI girlfriend do anything", which worked for sister app Bae Love, into roughly 17 languages. From 10-21 to 11-05 they ran it on Meta targeting South Asia, Southeast Asia, Central Asia and the Middle East, where installs are cheap, and bought 60k installs in two weeks. Apart from ads, there was no organic channel.

DHow you could follow it I wouldn't recommend copying an AI girlfriend shell. What's worth learning is the factory process: prove creatives on one app, reuse the playbook with a new package name and brand, translate one hook into a dozen-plus minor languages and push it to low-cost countries, and shut the whole batch down if it doesn't pay back in two weeks.

D grade: the category's top end sits in the tens of millions, the middle is crowded with this year's same-style paid-acquisition shells, and Michoo itself ran ads for two weeks, then stopped, with no proof it paid back. Staying alive also means risking Play delisting at any time. The playbook is worth understanding; following it isn't worth the effort. Few people know about it yet

Est. monthly revenue (Google Play) under $1KWe estimated this from its rank on country top-grossing charts and real install counts (it isn't on the grossing chart, so confidence is lower), calibrated against real revenue; typical error is 2–3x. Read it as a tier, not an exact number.

Data (updated daily) · first launch 2026-10-07 · 11-14 installs 62,603
Android · Total installs worldwide
20k40k60k11-0311-0511-0711-1211-1463k
Data 2026-11-03 → 2026-11-14 (12 days, 8 observations). Google's detail page only refreshes install counts every other day, so a flat day doesn't mean nobody installed.
Android · Ratings count (cumulative)
2829303111-1211-1311-1429
Data 2026-11-12 → 2026-11-14 (3 days, 3 observations)

Michoo is an Android-only AI girlfriend / roleplay chat app, listed in Oct 7. It had 74 installs in its first two weeks and reached 62,603 by Nov 12. The product itself has nothing new, and its store rating is 2.2. What drove its growth was one thing: a single suggestive ad, translated into about 17 languages and run on Meta for two weeks. Behind it is one person's app factory in Ha Long City, Vietnam, which already took another app to 1.44M installs using the same approach. Below: how it ran ads, how it collected money, why it stopped, and whether you should copy the approach.

An Oct 21 Hindi ad

In 2026 Oct 21, a Facebook page called "Michoo - AI Chat & Roleplay" started running a vertical video ad with one Hindi line: 「🍒अपनी AI गर्लफ्रेंड से वह सब करवाएं जो आप चाहते हैं।」, roughly "Make your AI girlfriend do whatever you want." Over the next week the same line appeared in Burmese, Uzbek, Bengali, Nepali, Persian, Sinhala, Tamil, Arabic, Russian, Turkish, Tagalog, Indonesian, Spanish, Portuguese, Thai and English, running on Facebook, Instagram, Audience Network and Threads (Meta Ad Library, searched by package name, 11-11).

18 days later, on Nov 8, a Google Play user named Andrew W left a review: "Thank you so much, it took 20 full seconds before it asked me to pay, I really appreciate it." That same day another user wrote: "The app shows an ad before it even opens... after that every page you tap and every option triggers another ad... after 324 lines of chat it can't remember the scenario it gave me."

Those two things pretty much are Michoo: cheap traffic, a suggestive hook, a paywall at 20 seconds, and an ad on every tap.

What it is: a generic AI girlfriend shell

The store page says "AI companion and roleplay": you can chat with companion, fantasy, mentor, anime and other personas, with "remembers context", "private", "instant replies" and "dark mode". Those lines would fit any similar app. The US store rates it 17+, labeled "sexual content, nudity". A review from Oct 28 put it more bluntly: "This app is full of explicit AI chat, and the ads push exactly that use."

Android only, not searchable on the App Store. The version stopped at 1.0.4 and has not been updated since Oct 22.

How installs grew: ads on, installs come in; ads off, growth drops by half

DateTotal installsWhat happened
10-070Launched
10-2274Almost nobody installed it in the first two weeks; 1.0.4 launch
10-25About 2,300Meta ads began running from 10-21, several hundred a day
11-01About 6,400
11-02About 22,600+16k in one day, the peak
11-0643,566
11-1262,6036 days, +19k, about 3,200 a day

Source: Google Play install counts. Daily increases are third-party AppstoreSpy estimates. Play updates its numbers in batches every few days, so single-day figures should be discounted.

From Nov 3 to 5, all 30 Michoo ads visible in the Meta Ad Library were paused. Each ran from Oct 21–29 to Nov 3–5, and none were live in Nov 11. After the ads stopped, installs kept climbing at about 3,200 a day. I couldn't trace where those came from: possibly a fresh batch of creatives the search doesn't show, Google app ads, or leftover Play "similar apps" recommendations.

Its chart placement is low: Entertainment free chart #89–190 in Thailand (11-03 to 11-07), #159–194 in Turkey, #180–185 in Poland, and #12 in the UK in Nov 174. It never charted on the overall or Top Grossing charts (Google Play charts). The ad languages target Myanmar, Nepal, Bangladesh, Sri Lanka, Uzbekistan, Iran and Pakistan, where installs are cheap, so most installs probably come from there (inference). The store doesn't publish installs by country, so this can't be verified.

60k installs produced only 28 ratings, 16 of them one-star. Lots of people installed it, but few stayed to rate. Most who rated left at the paywall.

Where the money comes from: a paywall at 20 seconds, plus an ad on every tap

Michoo collects money three ways at once:

  • Subscriptions and in-app purchases, store prices from $3.49 to $99.99. Per reviews, the paywall appears after just a few messages.
  • Ads, listed as containing ads in the store. Reviews say there are app-open ads and interstitials on page and option taps.
  • Impulse purchases. Someone who came in through "make her do anything" gets stopped right when they most want to keep chatting. That's the conversion point for this kind of app.

No figures are visible from outside on whether the business pays. One clue: Michoo's ads ran for only two weeks and were then all paused. For a team that was already buying and had creatives ready, the most likely explanation is that this account didn't clear its payback bar (inference).

Who's behind it: one person in Ha Long City, Vietnam, and a proven sister app

The Play developer is "Dog Games". The store page's "About the developer" names Truong Ngoc Anh, with an address in Ha Long City, Quang Ninh Province, Vietnam. The website is a Firebase-hosted page and the privacy policy sits on Google Docs. The account has 20 packages: AI cleaner, AI photo editor, AI home design, tarot, palm reading, AI translation, a browser, live scores, and small games like "dog grooming" and "cat brushing". AI Cleaner has 2.04M installs and AI Photo Maker 2.01M (Google Play, 11-12). This is a typical Vietnamese individual app factory: one person, many packages, growth bought through ads.

The closest match to Michoo is Bae Love: AI Girlfriend Chat from the same developer. It was listed in Jul 10 and went from 36k to 440k in the week of Aug 16–21. It now has 1.44M installs, grew another 63k in the last 7 days, and holds a 3.3 rating (Google Play, 11-12). It runs ads on Meta through two small Pages: one with 1,007 likes, 75 active ads and 605 in total; the other with 227 likes, 21 active ads and 111 in total. The copy is the same multilingual "🍒 100% free / Play with AI girls / Tell your AI girlfriend…" set (Meta Ad Library, 11-11). Several Bae Love ads still link to a dead store listing for another AI girlfriend app from the same developer: SoulBuzz: Chat with AI Friends.

According to AppBrain's records, SoulBuzz lived on Play for 44 days, was delisted in September, and had 11k installs in total. Michoo was listed in Oct 7. The "frnd" in its package name matches the "AI Friends" in SoulBuzz's subtitle, and the timing fits right after the delisting. Michoo looks like SoulBuzz relaunched under a new name (inference, with no official confirmation).

So Michoo isn't a new product. It's the second or third account from this factory: Bae Love is the proven main account, SoulBuzz was delisted, and Michoo was repackaged as "roleplay / multi-character" and tested next. Running several accounts in parallel is part of the playbook. Suggestive AI girlfriend apps can be treated as sexual content by Play at any time, and Play's generative AI policy doesn't allow sexually explicit generated content anyway.

Apart from ads, nothing else

Searching "Michoo AI" on TikTok returns only the unrelated "Mi Chico" dance template. On X in the last 60 days, Michoo and Bae Love each had 0 mentions. Reddit and media coverage are empty too. The only results online are auto-generated pages from APKPure, Softonic and AppBrain. Store search doesn't help either: the title is stuffed with "AI Chat & Roleplay", but with 2.2 stars and 28 ratings it doesn't rank for those terms.

So almost nobody knows about it, but the reason nobody knows is that it isn't worth talking about, not that an opportunity is hidden.

Who's competing

CompetitorsScale (Google Play total / last 7 days, 11-12)How Michoo differs
Character.AI70.9M / +155kTens of millions of user-made characters, community-driven; Michoo only has official preset personas
Talkie(MiniMax)60.9M / +117kBig studio, heavy paid acquisition, adds card collecting
PolyBuzz58.1M / +252kEqually suggestive, equally heavy on paid acquisition, but has 20M characters and a web version
Chai24.5M / +51kLong-running unfiltered chat
Replika20.3MLong-term emotional companionship, established brand
HiWaifu, Dokichat, Crushie, zeta810k – 9.51MAnime and romance focused, with content operations
Aiko AI, Kawaii AI, Emily, Amora, Mira400k – 2.35M, 7 days +10k – +120kThe same batch of paid-acquisition AI girlfriend shells launched between 4 and September this year. Same playbook as Michoo, one order of magnitude bigger
Bae Love (same developer)1.44M / +63kThe main account, same creatives
Grok companions, Meta AI charactersHugeBig-company features, and apart from Grok, almost none produce explicit content

The head is made of community products with tens of millions of installs, while the mid-tier is packed with copycat ad-driven shells that launched just this year. Michoo has no product differences from them, and in distribution it's just a smaller version of the same playbook.

If it were you

D. Understand this playbook, but don't follow it. The AI girlfriend category's top is in the tens of millions, and the mid-tier is a dozen-plus identical ad-driven shells that appeared within a year. Michoo is the developer's second attempt, ran paid ads for two weeks and stopped, and hasn't proven it can pay back. Staying alive long term depends on surviving removal at any moment, which is not something a follower can reliably get.

  • Who it's for: Utility app teams that have already made Meta paid acquisition work for an app, can make the numbers work in cheap countries, and have an existing multilingual creative pipeline. Treat it as another category to test for two weeks, not a product to build long term.
  • Who it's not for: Anyone trying to build an AI companion product with good word of mouth and retention (the users here come in from a borderline ad line and leave once they hit the paywall after 20 seconds, and the rating will be 2 stars); anyone unwilling to deal with content review and takedown risk; anyone who hasn't run ads and is counting on store search or organic social traffic (apart from ads, Michoo has no organic channel at all).
  • How to do it: If you do build it, don't make another generic English girlfriend shell. The part worth learning is Dog Games' factory process: prove the creatives and ad buying on one app, then reuse the same batch of creatives under a new package name and brand, with one test cycle every two weeks, and pull the whole batch if it misses payback. On the product side, the review gaps are clear: the paywall comes too early, the context gets forgotten, and the ads pop too often. Give enough free chat turns and do memory well, and you can at least lift the rating off 2 stars. Ratings are what suppress organic installs.
  • How to acquire users: Meta video ads, with one strong hook translated into a dozen-plus minor languages, targeted at South Asia, Southeast Asia, Central Asia, and the Middle East, where installs are cheap. Test about ten creatives per language for two weeks, check payback, and stop if it doesn't work (inference: this rhythm is read from when Michoo's ad runs started and stopped).
  • Deadline: There's no window to race. How long it runs depends on when Play tightens its review of this kind of app, which could happen at any time.
  • Signs I got this wrong: ① Michoo restarts ads on Meta and keeps them running for 30 straight days, with the number of live ads growing. That would mean the second app also paid back, and the playbook is easier to copy than it looks. ② From Nov 7 onward, roughly 3,200 installs a day keep coming in through December with no ads running. That would mean it has organic traffic, and a low rating isn't suppressing it. ③ The reverse: Bae Love or Michoo gets removed from Play, or Bae Love's live ad count is halved before year end. That would mean the door is closing.

Data and sources

  • Install counts, ratings, chart positions, store listing copy, developer info: Google Play (US and UK store listings, 11-11; installs up to 11-12). October's daily installs are AppstoreSpy estimates.
  • Verbatim reviews: Google Play US store reviews (10-28, 11-08).
  • Ads: Meta Ad Library, searched by package name com.dg.chatbot.frnd and "Michoo" (11-11); Bae Love and AI Cleaner homepages and live ad counts (11-11).
  • SoulBuzz's days listed, takedown date, and installs: AppBrain records.
  • Social and media: TikTok search for "Michoo AI", X mentions in the last 60 days, Reddit and news search, 11-11.
  • Competitor scale: Google Play install counts, 11-12.
  • Anything marked "inference" is inference. Ads were only checked on Meta; no results on Google or TikTok doesn't mean no spend there. There is no public data on revenue or payback.

Android store page ↗

Research notes: competitor table, acquisition evidence, full research report
Competitors
NameWhereLaunchedScaleHow it differs from this one
Character.AIAndroid2023-05(Android)Play 70.9M; 7-day +155kTens-of-millions UGC character community; Michoo only has official preset personas
Talkie(MiniMax)Android2023-06Play 60.9M; 7-day +117kBig company heavy paid acquisition + card collection mechanics; Michoo is a solo-operator template shell
PolyBuzzAndroid2023-06Play 58.1M; 7-day +252kEqually borderline and heavy on paid acquisition, but has a 20M character library and web SEO
ChaiAndroid2021-03Play 24.5MLong-running unfiltered chat platform; Michoo has no unique capability
ReplikaAndroid2017-09Play 20.3MLong-term emotional companionship brand; Michoo is a fast-moving consumer shell
HiWaifu / Dokichat / Crushie / zetaAndroid2022–2025810k–9.51MAnime and romance focused, with content operations; Michoo is a generic template
Aiko AI / Kawaii AI / Emily / Amora / MiraAndroid2026-04~2026-09400k–2.35MA copycat AI girlfriend shell buying the same traffic in the same period, same playbook; Michoo is an order of magnitude smaller
Bae Love (same developer)Android2026-07-10Play 1.44M, 7 days +63k, 3.3 starsThe main account running the same Meta creatives; Michoo repackaged as roleplay / multi-character serves as the second account
SoulBuzz (same developer, removed)Android2026-0811k installs, listed for 44 days, removed in September (AppBrain)Michoo's predecessor, likely renamed and republished after being removed for policy violations
ChatGPT / Grok companions / Meta AI charactersBig-company features2024–2025HugeBig companies largely don't make explicit content (Grok is the exception); Michoo sells borderline content
Where users come from
  • Confirmed Meta ads (Facebook / Instagram / Audience Network / Threads): Searching the Meta Ad Library by package name, the first 30 results are all video ads from the Pages "Michoo" / "Michoo - AI Chat & Roleplay". Ads started 10-21 and ran through 10-29 11-03 to 11-05, all stopped. The copy "🍒 Tell your AI girlfriend to do whatever you want" appears in about 17 languages: Burmese, Uzbek, Bengali, Nepali, Persian, Sinhala, Tamil, Hindi, Arabic, Russian, Turkish, Tagalog, Indonesian, Spanish, Portuguese, Thai, and English. The install curve matches the start and stop dates.
  • Confirmed Reuses creatives and ad buying from sibling app Bae Love: In the metaads library, two low-like Pages for Bae Love (1,007 likes / 227 likes) have 75 + 21 live ads, with 716 historical ads, all using the same multilingual 🍒 copy. Some still link to the removed SoulBuzz. The AI Cleaner Page has 106 live ads.
  • Inference About 3,200 a day of residual installs after ads stopped (new creatives / Google UAC / Play recommendations): 11-07 to 11-12 +19,037. No live Meta ads were found in the same period. The googleads and tiktokads libraries both show no_ads for the sibling apps.
  • Inference TikTok / X / Reddit / media: A TikHub search for "Michoo AI" returned 0 related results, x-radar shows 0 mentions, and no media coverage was found. These channels are essentially absent.

Michoo: AI Chat & Roleplay (com.dg.chatbot.frnd) research report

Research date 2026-11-11 (data card refreshed to 2026-11-12). Data sources: our googleplay / metaads / tiktokads / googleads / appstore libraries (agent-im server), Google Play store pages, the Meta Ad Library web pages (fetched via Jina), Jina / Exa web search, TikHub TikTok search, x-radar.

One line

Michoo is the AI girlfriend / roleplay chat shell of a solo developer in Ha Long City, Vietnam (Truong Ngoc Anh, publicly known as "Dog Games"), and is essentially the removed SoulBuzz republished under a new name. It didn't grow as a product: most of its 60k installs came from Meta ads bought in the 10-21 to 11-05 two-week window. The ads use the same borderline "🍒 Make your AI girlfriend do anything you want" video, translated into about 17 languages and targeted at Southeast Asia, South Asia, the Middle East, and Central Asia, where installs are cheap. Store rating: 2.2. The real thing to study is the playbook behind it: the same developer used the same method to take Bae Love to 1.44M installs. Also, the recall reason's claim that it "only ran ads in 1 countries" is wrong.

1. Who they are

  • Developer: Google Play lists "Dog Games", but the "About the developer" section names Truong Ngoc Anh, with an address in Ha Long City, Quang Ninh Province, Vietnam (To 3, Khu 6A, Hong Hai, Thanh pho Ha Long). Email [email protected], website ezp-meme-battle.web.app (a Firebase hosting page), privacy policy on Google Docs [Play store page, gl=gb, 2026-11-11]. This is a textbook Vietnamese solo app factory.
  • Portfolio: The developer page lists 20 packages, all com.dg.* / com.ai.*: AI Cleaner, AI Photo Maker, AI Home, Bae Love, EZchat (com.dg.assist.chatbot.ai), Tarot AI, AI Palm, Planty AI, AI Food, AI Translate, Veil browser, Live Score, Speaker Cleaner, Beat Maker, plus small games like Dog Barking, Brush The Cat, and Trim The Dog [Play developer page]. The 5 apps in the library total 5.59M installs: AI Cleaner 2.04M, AI Photo Maker 2.01M, Bae Love 1.44M, Michoo 63k, AI Home 37k [googleplay library developer / app_profile, 11-12].
  • Predecessor SoulBuzz: com.dg.girlfriend.dreamgirl, "SoulBuzz: Chat with AI Friends". AppBrain records it as "on Google Play for 44 days, removed in September 2026, 11k installs" [appbrain.com search summary]. Its Facebook Page (61593771231676) ran video ads for "Play with AI girls" [Facebook search results]. Michoo launched 10-07 with the package name chatbot.frnd (friend), and SoulBuzz's subtitle "Chat with AI Friends" lines up with it. The timing comes right after SoulBuzz was removed. "Renamed and republished" is an inference, with no official statement. There's also corroborating evidence: in Bae Love's Meta ad copy from October, a dead link to SoulBuzz's com.dg.girlfriend.dreamgirl is still present [metaads library ad table], showing that creatives and copy were reused in bulk across several apps.
  • Product: AI character chat + roleplay, with optional personas such as companion, fantasy, mentor, and anime. The description is all generic copy ("remembers context", "private", "instant replies", "dark mode"). PEGI 16, tagged "Sex" and "Users Interact" (the recall said the US rating is Mature 17+, with sexual content / nudity) [Play store page]. Android only; not found on iOS [appstore library].
  • Monetization: "Contains ads" + in-app purchases from $3.49–$99.99 (appfigures lists subscriptions + in-app purchases + ads) [googleplay library app_snap; appfigures metadata page]. Nothing has changed since version 1.0.4,10-22.

2. Growth data

Install curve (googleplay library app_day / app_day_ext; October data is AppstoreSpy estimates):

DateTotal installsNotes
10-070Launched
10-1928Almost no installs in the first two weeks
10-22741.0.4 launch (10-22)
10-252,25610-23 to 10-25 about 700 a day, Meta ads start 10-21
10-292,576
11-016,436
11-0222,625Single day +16,189, scaling up
11-0532,116
11-0643,566+11,450
11-1262,6036 days +19,037, about 3,200 a day
  • 7 days +30,487 (data card). Only 28 ratings, averaging 2.2: 1 stars 16, 3 stars 5, 5 stars 5 [app_snap 11-12]. Lots of installs, almost no ratings, and most of those are still 1 stars. That's the typical look of paid-acquisition users: they install, hit the paywall, and leave.
  • Chart presence: only tail spots on Top Free in entertainment across several stores: Thailand 89–190 (11-03 to 11-07), Turkey 159–194, Poland 180–185, UK 174 (11-12) [googleplay lib chart_rank]. Never charted on the overall or Top Grossing charts. The recall claim that it's "only running in 1 countries" is wrong: n_countries_now = 1 only means that on 11-12 it was only on the UK chart. In earlier days it charted in Thailand, Turkey and Poland too, and ads ran to dozens of countries.
  • Most installs likely come from Myanmar, Nepal, Bangladesh, Sri Lanka, Uzbekistan, Iran, Pakistan and similar places where the ad language matches but our 30-country chart coverage doesn't reach. This can't be verified directly.

3. Competitors

ProductWhereLaunchedScale (Play lifetime / 7 days, googleplay lib 11-12)How it differs from Michoo
Character.AIAndroid / iOS / web2022 / 2023-0570.9M / +155kTop community-type product with tens of millions of UGC characters; Michoo only has official preset personas, no community
Talkie(MiniMax)Android / iOS2023-0660.9M / +117kBig-company heavy paid acquisition + card-collection mechanics; Michoo is a solo small shop
PolyBuzzAndroid / iOS / web2023-0658.1M / +252kEqually suggestive, equally paid-acquisition heavy, but has a 20M character library and a web version for SEO
ChaiAndroid / iOS2021-0324.5M / +51kLong-running unfiltered chat app; Michoo has no unique capabilities
ReplikaAndroid / iOS201720.3MFocused on long-term emotional companionship, established brand; Michoo is a fast-consumer shell
HiWaifu / Dokichat / Crushie / zetaAndroid / iOS2022–2025810k – 9.51MAnime and romance focused, with content operations; Michoo is a generic template
Aiko AI / Kawaii AI / Emily / Amora / MiraAndroid2026-04 ~ 2026-09400k – 2.35M, 7 days +10k – +120kSame batch of paid-acquisition AI girlfriend shells this year, same approach as Michoo; Michoo is an order of magnitude smaller
Bae Love (same developer)Android2026-07-101.44M / +63k, rating 3.3Sibling app, same Meta creatives; Michoo just repackaged as "roleplay / multi-character", effectively a second account
SoulBuzz (same developer, delisted)Android2026-08 ~ 2026-0911k (AppBrain)Predecessor of Michoo, likely delisted for violations
ChatGPT / Grok companions / Meta AI charactersBig-company features2024–2025HugeDoesn't do explicit content (Grok excepted); Michoo sells exactly that suggestiveness

Verdict: the category is extremely crowded. The top are tens-of-millions community products, and the middle is packed with paid-acquisition shells launched this year. Michoo is no different from them in product.

4. How users came in (growth levers)

  1. Meta ads (confirmed, main channel). Searching the Meta Ad Library by package name com.dg.chatbot.frnd returns 30 ads on the first screen, all video ads run by the Pages "Michoo" / "Michoo - AI Chat & Roleplay". Ad start dates cluster between 10-21 and 10-29, and everything stopped between 11-03 and 11-05 (none were running when 11-11 was scraped) [facebook.com/ads/library, keyword=com.dg.chatbot.frnd and keyword=Michoo].
    • The copy is the same line, "🍒 Tell your AI girlfriend to do whatever you want", translated into Burmese, Uzbek, Bengali, Nepali, Persian, Sinhala, Tamil, Hindi, Arabic, Russian, Turkish, Tagalog, Indonesian, Spanish, Portuguese, Thai and English, about 17 languages in all.
    • Placements: Facebook, Instagram, Audience Network, Threads.
    • Matches the curve: 10-23 started at a few hundred a day, 11-02 single day +16k; after ads stopped, Play installs kept rising to 11-06 (store updates lag).
  2. The whole playbook was copied from Bae Love (confirmed). Bae Love runs ads through two Pages: "Bae Love - AI Girlfriend Chat" with 1,007 likes, 75 active ads, 605 total, earliest seen 08-26; and "Bae Love AI Chat" with 227 likes, 21 active ads, 111 total. All video, copy is the same multilingual "🍒 100% free / Play with AI girls / Tell your AI girlfriend…" set [metaads lib advertiser, 11-11]. Bae Love went from 08-16k to 08-21k in a week from 36 to 440 (single days 100k+) [app_day_ext], matching the time it started Meta ads. The same developer's AI Cleaner works the same way: Page with 1,307 likes, 106 active ads, 862 total. The read: Dog Games' growth rests entirely on "low-like Pages + a large volume of multilingual video creatives + direct Play links".
  3. Google Ads / TikTok Ads: not findable in the libraries. tiktokads returns no_ads for Bae Love (that library only covers Europe); googleads returns no_ads for both Bae Love and AI Cleaner [googleads lib watch table]. Michoo isn't in either library's list. Ads for 11-07 to 11-12 stopped, yet there's still roughly 3,200 a day in incremental installs. Possible sources: new Meta creatives not found by package-name search, Google UAC, or leftover Play "similar apps" recommendations. Not verified.
  4. Creators / influencers / community: none. Searching "Michoo AI" on TikTok returns only unrelated "Mi Chico" dance templates [TikHub]; on X, Michoo and Bae Love each get 60 mentions in the last 0 days [x-radar]; no Reddit or press coverage either. What turns up online is only auto-generated pages from AppBrain, appfigures, Softonic, CNET Download and APKPure.
  5. Store search / ASO: little effect. The title stuffs in "AI Chat & Roleplay", but with 2.2 stars and 28 ratings it doesn't rank for those terms.

5. Why it gained traction

  • Paid acquisition, not product. The install curve lines up with the Meta ad window: 10-21 started, 10-23 volume ramped, 11-02 peaked, 11-05 ads stopped. In the 15 days before that there were only 74 installs in total.
  • Cheap traffic paired with suggestive creatives. The "make your AI girlfriend do anything" hook is pushed into South Asia, Southeast Asia, the Middle East and Central Asia, low-CPI markets, in 17 languages at once. Creatives can be reused straight from Bae Love, so marginal cost is very low.
  • Keep spending on a new account. SoulBuzz was delisted, then Michoo was relaunched under a new name right away, while Bae Love kept running. Running several accounts in parallel spreads the risk of delisting; apps like this are easily flagged by Google as sexual or misleading content.
  • Where the money comes back from: three streams at once: ad monetization ("contains ads"), subscriptions / in-app purchases ($3.49–$99.99), and impulse payments in low-price markets. A 2.2 rating suggests very poor retention, so payback depends on paywall conversion in the first few days plus ads. Whether the business pencils out has no public data; Michoo stopping everything after two weeks looks more like a test that didn't clear the ROI bar (inference).

6. Differentiation

No product differentiation. Persona, chat, memory and privacy selling points are shared by every peer. The only real difference is distribution: a one-person Vietnamese studio uses very cheap creatives, translates the same suggestive videos into a dozen-plus minor languages, targets long-tail low-price countries that big companies and Chinese going-global teams don't prioritize, and hedges delisting risk with multiple shell apps. This playbook worked for Bae Love (1.44M). Michoo is currently just a test run on a second account and hasn't proven itself yet.

7. Visibility and copycats

  • Visibility: barely known. No press coverage, 0 mentions on X, no creator content on TikTok, and revenue has never been disclosed.
  • Copycats: nobody has copied Michoo. Instead, it is itself one of many near-identical shells in this category this year (Aiko, Kawaii AI, Emily, Amora, Mira, etc.). The developer is even cloning themselves (SoulBuzz → Michoo, running alongside Bae Love).

8. What to borrow

  • Get one app's creatives and bidding model working first (Bae Love), then open a second and third account under new package names and brands, reusing the same creatives, both to spread delisting risk and to test other packaging (roleplay and girlfriend).
  • One strong hook line, translated into a dozen-plus minor languages, pushed into low-CPI countries for volume, while building install counts and exposure from the "entertainment chart tail".
  • Two-week cycles: if ROI doesn't work, pull the whole batch and don't linger.
  • Risks: suggestive content gets delisted from Play (SoulBuzz is the warning), a low rating suppresses organic installs, and it can only rely on paid acquisition long-term.

9. Uncertain

  • Where the 11-07+ daily incremental installs after 3,000 come from (new creatives, Google UAC, or organic recommendations) is unverified.
  • Country distribution of installs: inferred from ad languages to be mainly South Asia, Southeast Asia and the Middle East, which our chart database doesn't cover, so it can't be verified.
  • The exact date and reason for SoulBuzz's delisting (the recall says 09-26), and "Michoo is SoulBuzz relaunched under a new name", are inferences.
  • The recall says Bae Love "charted on the Lebanon Top Free 21", but Lebanon isn't in our 30-country coverage, so it's unverified.
  • Revenue and ROI: no public data at all.

Sources

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