11 days, 55k installs, not a single review: how a Vietnamese app factory is reselling "AI car modification"

Car Modify: AI Car Designer launched in Oct 26 and reached 55k installs in 11 days. It ranks on the free Art & Design charts in 19 countries, #75 in the US. Its product is nearly identical to ModsTune, which proved the model three months earlier. It comes from a Vietnamese studio that has already shipped a third package on the same theme. What's worth studying is its distribution playbook. The product itself has little to learn from.

A Reel with a "Download Free" button

In Oct 27, the day after launch, a Facebook page called "Car Modify: AI Car Designer" posted a Reel. The caption read "Transform your car with AI Car Designer! Upload a photo…Try it now!" with a "Download Free" button below. The video got 5,300 views and 282 reactions, and the page itself had only 4 followers (Facebook page, read 11-06).

A page with 4 followers getting thousands of views on one video usually means it was paid ads. Two days later, on Oct 29, its Google Play installs were only 46, yet it was already #67 on the free Art & Design chart in Vietnam, #80 in Nigeria and South Africa, and #113 in Turkey.

What it is

Upload a photo of your car and AI generates modification previews: paint changes (matte, pearl, chrome), wraps and decals, wheel swaps, body kits and spoilers, lowering the car, plus removing license plates and turning the preview into a short video. The last line of the store listing is "Preview and refine choices before going to the shop": see the result before you spend money.

Ads only, no in-app purchases. Version 0.0.2, no updates since launch, and as of 11-06 it still had no ratings and no reviews (Google Play, 11-06).

The numbers: charting in cheap countries first, then the US

Google Play installs only update every few days, so one day's numbers mean little. Look at the trend. The order is clear: Vietnam, Nigeria, South Africa, Turkey, Thailand, Indonesia, and the Philippines first, then Canada, the US, Australia, and Taiwan. This is the classic paid-acquisition rhythm: test creatives and build volume in cheap countries, then spend money in expensive ones once it works (inferred, the Meta ad library returned no ad records for it).

There's another point: the "Art & Design" category has very weak competition. A few thousand installs a day can put an app in the US top 100, and the chart position itself brings some organic exposure.

Who's behind it: one email, three developer accounts

The developer is listed as MEDICI TECHNOLOGY GROUP. Searching Play for its support email [email protected] turns up at least three developer accounts run by the same team:

There are also anime wallpapers, phone cleaners, ringtones, browsers, PDF readers, and a stream of other ad-monetized utility apps. The website midasstudio.app describes itself as "10 Apps live on Play" and "ASO & growth," and it's hiring remote "User Acquisition" staff and a product manager in Hanoi. The address registered on Play is in Ho Chi Minh City (website and Play listing, 11-06). It's essentially an app factory that lives on traffic.

In six weeks, the same theme shipped three packages, with the second and third only a day apart. All three packages occupy keyword slots like "ai car designer" and "car modify" at once. If one is dragged down by bad reviews or removed, the other two keep running.

Why the first package only has 1.3 stars

The studio's first app was Midas: AI Car Modification, launched in September, with 157k installs and a rating of 1.3 (344 ratings). The three most "helpful"-voted negative reviews on the store page (Google Play, 11-06):

"It doesn't even modify the way you choose. It's not even your car. I uploaded a photo of my car, picked the settings and colors, and it produced a completely different car with the wrong color." (10-13,10 people found this helpful)

"Terrible, way too many ads. You have to watch ads before it generates anything, and then it swapped my car for a different one. Garbage. Free AI tools are better than this." (10-06)

"Watched six or seven long ads, and then it gave me some random image from the internet and said it was my modified car. Don't waste your time." (10-12)

The complaints boil down to two things: too many ads, and the output isn't your car. The second is an inherent flaw of this category: AI image editing easily "changes the car into a different one partway through." The two newer packages dropped the paywall, relying only on ads and free access, and restarted under new developer accounts. Whether the generation quality can be fixed and whether the ads will stay just as heavy will only be clear once the new packages collect reviews.

The demand is real, and someone else proved it first

The leader in this niche is ModsTune – AI Car Designer, launched in Jul 23. In three months it reached 624k installs, 30k ratings, and a 4.55 rating, ranked #6 on the Art & Design Top Grossing chart in Mexico, and earned money from in-app purchases of $9.99–29.99 (Google Play, 11-06). Its reviews in Mexico are mostly positive. The negative ones almost all say the same thing:

"A paid app, and you can't do anything without paying." "Nothing works without adding a credit card." "A week for 11,200, no way." (Mexico store reviews, October)

The content side is hot too. On TikTok, a Spanish-language video posted in Oct 24 showed people how to use CapCut's "JDM filter" to turn their own car into a lowered, sporty-wheeled, big-spoiler build, with before-and-after shots. It got 380k views (TikTok, read 11-06). There are also YouTube tutorials on "how to make the viral AI car-modification video." Lots of people want to see what their own car would look like modified, and most of them don't want to pay for it.

Who it's competing with

At least 8 similar apps charting on Play; searching "AI car designer" on iOS returns a full page of clones, and several new ones have launched in the last 2026 years (App Store search, 11-06). This category has plenty of products already. Whoever gets people in the door at the lowest cost wins the volume.

Barely anyone is watching

No press coverage, almost nobody on X has mentioned "AI car modification app" in the last 30 days, there's no public revenue, and no creator videos promoting it turn up. Few people know it, but this lane is already crowded.

If it were you

C: Worth learning from, not worth chasing. It gets its volume from distribution: one studio, three accounts, buying cheap-country traffic to climb the category charts first, then expanding into high-value countries and monetizing with ads. Building another "AI car modification" app to copy this is the N+1th one, and you'd be competing on who can buy traffic more cheaply.

Who it suits: teams that already run an ad-monetization tool stack, know how to adjust Meta budgets by country, and can work out what a user is worth in ad revenue down to the cent. For you, this is just another theme, and the reskin cost is very low.

Who it doesn't suit: indie devs hoping to make money from the "AI car modification" idea itself. There's no feature moat, the underlying image models are off the shelf, and free CapCut templates and ChatGPT can do the same job. The leader, ModsTune, already has 30k ratings.

How to do it: if you do it, go after the part the others do badly. First, the output has to be the user's own car. That's where clone reviews complain most, so get model, body lines and angle preservation solid before adding bells and whistles. Second, connect to real products: recommend wheels, body kits and wraps that can actually be bought for that model, link to local body shops or e-commerce, and earn commission rather than ads. The "check before you go to the body shop" line on store pages is only really delivered by a handful of players like MOTORMIA.

How to get users: the creative is vertical before-and-after short video, one photo of the original car, then a clip of the finished result. This format grows views on TikTok and Reels by itself. Test first in Spanish-speaking markets (Mexico, Colombia, ModsTune's home turf, where reviewers complain about the price) with a free-plus-ads version. If you want to charge, partner with car-modding communities and body shops rather than pushing weekly subscriptions.

Timing: interest in this theme follows short-video trends, and it is currently near the top. If you want to test, have a usable version within a month and run two to three weeks on a small budget. If after two weeks ad revenue per user is still below acquisition cost, stop.

What would show I'm wrong: (1) if Car Modify and CarDes keep growing installs a month or two from now with ratings above 4, the ads-only approach is working and the window to follow is bigger than I think; (2) conversely, if these two packages stop updating and drop off the charts quickly, or their ratings fall below 2, this lane is a passing fad; (3) if any "real parts / body shop" app shows visible revenue, the second path works and is worth serious effort.