Three empty Facebook pages, a Vietnamese app factory: how Mervious bought 1.93M installs in three and a half months

Why Mervious: AI Photo & Video Art (CrocoTech) was able to take off

Data card: play_com.aiphotovideo.generator.imagetovideo; research date 2026-11-06.

Bottom line

Mervious is one of dozens of "AI portrait / image-to-video / viral template" wrapper apps on Google Play, with no product differentiation. It gained traction because it sits in a network of Vietnamese app factories: multiple developer accounts publish in bulk through the same pipeline and build sites by project number. The factory used several rounds of Facebook ad spend, and the first round pushed it into India's Photography Top Free top ten; afterward it rode organic traffic from category charts in India, the Philippines and the US. Monetization relies on "watch a pile of ads first, then a $19.99–$39.49 paywall". Funkedenic, a account in the same network, has already launched nearly identical Miuci AI and Mistmere.

1. Data (googleplay library)

• Installs: 1,929,018 lifetime worldwide, +190,946 over the last 7 days (app_profile, 11-06). 07-20 listed, about 3.5 months.
• Install curve (app_day_ext, AppstoreSpy estimates; single-day values are averages between two observations, so read the shape only):
  • The week of 08-10 it had only 16k cumulative; the week of 08-17 it added +136k (about 81k/day on average). This was the first spike.
  • From 09-07 to 09-20, two weeks, +447k (about 09-16 averages 156k/day).
  • Two spikes, on 10-04 and 10-07 (averaging 142k and 173k/day), pushed cumulative past 1M.
  • One spike on 10-23 (averaging about 181k/day) took cumulative from 1.45M to 1.63M.
  • Our own pulls: 10-29 to 11-06:10-31 +124k (two days combined), 11-03 +48k, and after that about 0.9–10k a day (app_day).
  • Monthly totals: August about 228k, September about 571k, October about 1.05M, first 6 days of November about 66k.
  • The curve is staircase-shaped: a spike, a fall-off, then another push two or three weeks later. That's the classic shape of budgets released in batches, not the smooth compounding of word of mouth.
• Charts (chart_rank, from 10-29):
  • Photo & Video Top Free in India: 10-29 #13, then slid steadily to #63 by 11-04, and back to #42 on 11-06. A search snippet from Sensor Tower's public page shows it at #7 on the India photo chart on 10-27 (Jina search summary; the original page needs login and was not opened directly).
  • US Photo & Video Top Free: 108 → 61 (11-02/03) → 74; Philippines 54 → 36; best in Italy was #62.
  • Today it's on the Photo & Video Top Free chart in 15 countries. It has never charted on any Top Grossing list.
• Ratings: 8,955 total, averaging 4.58, with rating count about 0.46% of installs. Adds 60–90 ratings a day. Only 15 have written reviews (written_reviews in app_snap). The star distribution is odd: the 2-star count stayed at 0,3 until 11-04, then went from 53 to 0,1, and the 690- and 564-star counts keep swapping places (histogram in app_snap). Miuci, on the same network, also saw 1 and 2-star counts drop suddenly. The cause is unknown. Google may have purged a batch of reviews, or the ratings may be manipulated. Not verified.
• Monetization: ad-supported (ad_supported = True), in-app purchases $19.99–$39.49 each (app_snap). That price range is far above Miuci's on the same network ($4.99–$17.99), which suggests weekly or annual subscriptions.
• The most-liked reviews on the store page show what the product feels like:
  • "Too many ads, I tapped 'next' for half an hour, and in the end it wants a subscription" (Nitesh Kumar, 09-28,47 people found this helpful);
  • "Can't make the video from the ads, what it makes looks nothing like the photo I uploaded" (Ryan He, 08-12,52 people found this helpful);
  • There's also a positive review with 192 upvotes: "the results look natural" (08-24). (Play store page, Jina scrape)

2. Developer: a Vietnamese app factory network

• The "About the developer" section lists an individual name, VU THI THEM, email [email protected], address Hong Khe, Binh Giang, Hai Duong, Vietnam (Play store page). CrocoTech is just the display name; it's a Vietnamese individual developer account.
• Contact email [email protected]. The website is an empty Firebase site, aap878-ai-photo-crocco.web.app, and the privacy policy sits on Google Sites.
• The same account also has Hairisi: AI Hairstyle Try on (listed 08-08, 674k installs, +57k over 7 days), website aap877-ai-hair.web.app (googleplay library app).
• Searching the library for the signature "aso.<name>@gmail.com email + project-number.web.app website" turns up a whole network (googleplay library app/app_profile, 11-06):

Funkedenic's Miuci was listed one day before Mervious, with nearly identical package names (com.aiphoto.aivideo.generator.photofilter vs com.aiphotovideo.generator.imagetovideo) and consecutive website IDs (aap875 / aap877 / aap878). It's a strong inference that these two accounts are the same publisher or the same pipeline: one product, different names and icons, uploaded under different accounts. Whether the whole network is one company or a circle sharing an ASO or ad vendor is unverified.

3. Where users come from

1. Facebook ads (direct evidence, number of ads unknown): Facebook has at least 3 Pages named after Mervious, all posting short videos with "Download now / Try here" call-to-action buttons, the format of app install creatives (Jina search and scrape):
   • "Mervious" (id 61590623700402): 10-22 posted "Top 1 app AI photo to create viral video for you", 1.8K views;
   • "Mervious: AI Photo & Video Art" (id 61593442674095): 10-07 posted the same copy;
   • "Mervious - AI Art" (id 61590683916560): "Create a viral video with this app". These Page IDs all come from a new number range, with no visible followers or organic content. They are ad shells. The creative posting dates line up with the install spikes: 10-07 creative against the 10-07 spike (averaging 173k/day), and 10-22 creative against the 10-23 spike (averaging 181k/day). Our Meta Ad Library (we only found advertisers in keyword top results for US/GB/DE/JP), Google Ads Transparency and TikTok Ad Library did not track it, so the number of ads and the spend are unknown.
2. Reviews corroborate ads: "no options to make videos seen in ads" (Play store review, 08-12). Someone came in from an ad within three weeks of listing.
3. Organic traffic from charting and search after the push:
   • The India Photo & Video Top Free chart has a low bar. On 11-06, #6 Artify AI added only 7k worldwide over its 166 days, and #13 Lumpix added 237k (v_chart). Mervious adding over 100k a week was enough to hold a spot in the top 15.
   • The package name imagetovideo and the title "AI Photo & Video Art" are keyword stuffing, and the email username even includes "aso".
4. Copycat tutorials in India (a result, not a cause): YouTube has several Hindi "how to use Mervious ai app (kaise use kare)" videos (channels @webmindofficial, @atishapp, etc.), one with a Telegram download group t.me/Tusharapp1. These small channels chase whatever trending app is hot, with tiny view counts (one had only 11 views after 13 days, per the Jina search summary). They confirm the app was visible in India, but not that it was a major acquisition channel.
5. No other spread: X (x-radar, 60 days) shows zero mentions; a TikTok search for "mervious" returns nothing (TikHub); no media coverage. Searchable mentions are limited to automatic listings on APK sites like Softonic, Uptodown, APKPure and AppBrain.

4. Competitors

There are a great many Android apps doing the same thing (turning selfies into AI portraits with templates, or turning photos into short videos). Miuci's report from the same network counted 32 apps listed within 180 days, with names containing AI and photo/video/image, that have passed a million installs, totaling 125M installs. The main groups are below (googleplay library, 11-06):

Verdict: crowded market, and many of the competitors are "siblings" from the same network. Mervious ranks mid-pack by scale.

5. Differentiation

On the product level there's no visible differentiation: the feature list (portraits, ID photos, couple/family/pet photos, Old Money style, football themes, image-to-video) is standard for this type of app, the underlying model isn't disclosed, and reviews complain "the results don't match the ads." Compared with similar apps, only three things actually differ:

• it targets India and the Philippines first;
• its creatives push "Top 1 app for viral videos";
• its paywall is priced higher ($19.99–$39.49, while its twin Miuci is $4.99–$17.99).

So it's getting traction not because of differentiation but because of the factory network's ad-buying cadence, plus organic traffic from the low-barrier category charts.

VI. Visibility

Few people know about it yet. Revenue isn't public, it hasn't charted on any top grossing list, X, TikTok and media don't discuss it, and only small Indian YouTube channels with tutorials and APK sites index it. But the idea itself is a crowded red ocean, and people are already copying it: the same network has shipped Mistmere, a reskinned new app.

7. What's worth borrowing

• Matrix publishing: multiple individual developer accounts publish the same product (Mervious / Miuci / Mistmere), with empty Firebase sites named "project-number-appname.web.app" and privacy policies hosted on Google Sites. If one app gets pulled or creatives fatigue, the overall impact is small.
• Wave-based ad buying: a concentrated push every two to three weeks, each one driving daily installs above 150k to hit the top of category charts in low-barrier markets (India, the Philippines), then the next wave once it fades.
• Disposable ad pages: one product runs several Facebook Pages posting short videos with FOMO copy like "Top 1 app / viral video."
• Monetization funnel: heavy interstitial ads first, so free users also cover some cost, then high-priced subscriptions. The cost is one-star reviews complaining about "too many ads" and "doesn't match the ads."

8. What's still uncertain

• How many ad creatives ran, how much was spent, how spend was split across countries, and whether Google UAC was also used: none of the three ad libraries tracked it.
• Whether it pays back: no top grossing rank, and ad revenue isn't visible.
• Whether this network is one company or a circle of shared ASO / ad-buying vendors: developer site numbers are sequential and email formats match, but there's no direct evidence.
• Why the rating distribution is abnormal (2-star at 0 and 3-star suddenly zeroed).
• Whether the Indian YouTube tutorials are organic or paid for by the studio.