An AI fantasy video app that hasn't updated once, and it's charting in Korea's top grossing

Tomi: Mythora Studio is an Android app that turns a photo into a fairy, dragon or magic forest video. It launched in Oct 3 and has never had an update. A month later it had 18k installs and sat around #130 in Korea's Google Play entertainment top grossing chart, and from November it also charted in Japan. It has no organic buzz and a pile of bad reviews, and at least a dozen identical apps exist. The interesting part isn't this app but how these template shells make money, and whether you should get in.

"Promised a refund when cancelled"

In Oct 23, a Korean user left a review on Tomi's store page: "Scammers, give me my money back. I already cancelled the recurring payment, and you said cancelling would refund me, so isn't this a scam too?" Four days later another: "Paid, but the points never arrived, and there's no contact info in the app. Don't use it." The next day: "Thought you'd give free coins, but everything costs money ㅠ" The US store review was blunter, Oct 17: "Misleading ad. Repackaged porn app with E rating" (misleading ad, a repackaged porn app, yet rated for everyone).

By Nov 6, its store page had 38 reviews, not enough to show a star rating, and nearly all of them looked like these. That same day it ranked #139 in Korea's Google Play entertainment top grossing and #169 in Japan's entertainment top grossing. Reviews full of complaints, yet the chart keeps charging. That's normal for this kind of app.

What it is

The store page screenshots show lavender castles, clouds and flowers, and a blonde girl with sparkles and wings. The titles read "Transform Your Photos" and "Turn portraits into radiant fantasy creations." The website tomi.buzz lists the styles: Fairy Glow (flower crowns, glowing wings), Enchanted Forest (fireflies and glowing mushrooms), Dragon Dusk (a dragon over sunset ridges), Sky Castle, Starlit Mage, Deep Sea Realm. Each template takes one photo and generates a 9:16 short video. You can also type a sentence to generate one.

There's no in-house model. The likely setup is an off-the-shelf image-to-video API wrapped in templates. The developer is Lustre Collective LLC, a US LLC name, with a personal Gmail as the contact. The site went live in Oct 4 as one landing page plus a privacy policy and terms. The version number has been 1.0.0 since launch, and the "What's new" note still says "Welcome to the first official release of Tomi!" Another app under the same developer account, "VitraVault Quake Lab," has nothing to do with it by package name or product, and looks like an account reserved in advance.

The description was translated into Korean but the screenshots are still English, so Koreans are shown an English "fantasy portrait" poster.

Where the money comes from: you pay as soon as you install

It charges a subscription plus credit packs, priced at $4.99 to $399.99 per item on Google Play. There's no free allowance, and several reviews complain about it. Credits show on the profile page, and the sample account in the screenshot reads "Credits 10000 · Get credits."

With no free uses, every installer either pays on the spot or uninstalls. That brings day-one revenue and gets ad spend back quickly. The cost is bad reviews, refund disputes and the "misleading ad" accusations. Installs are small: Oct 19 530, Oct 29 8,746, Nov 6 18,616, and 7 new installs in the last 9,870 days (Google Play install counts). On Korea's and Japan's entertainment top grossing charts there's little competition from utility apps, so around #130 the bar is low, and 1–20k installs with a high enough paying rate gets you onto the chart. How much it earns per day isn't visible from outside.

Where users come from: bought, and the creatives may be borderline

No organic source turns up. The official YouTube channel "Mythora Studio" has 5 subscribers and 2 dragon-themed Shorts, with 141 and 34 views. Searching TikTok for "Mythora Studio" returns an unrelated AI short-drama account. On X there has been no discussion in the last two months, and no media has written about it. The only hits are automatic listings from APK download sites.

The clues all point to paid acquisition:

• The privacy policy names an "attribution partner" that reads the advertising ID to "tell us how you found Tomi" and "which campaigns bring new users";
• Install counts jump in steps: Oct 20 +2,602, Nov 1 +3,121, Nov 3 +4,396, with almost no movement in between, like batches being bought;
• That "Misleading ad" review.

Which network it buys from (Meta, Google or an ad network), and which creatives, can't be found in public ad libraries. "The creatives skirt the line" rests on a single review (inferred). The terms mention "where offered" third-party web payments; whether that's live is unverified.

Who it's competing with

Within this batch Tomi is the small one. Its only difference is the "fantasy" skin, and that skin isn't why it's working. The closest peers are bigger and launched earlier, and one team launched four accounts at once.

What this business actually is

It's a bidding business, not a product business: plug in a video model API, wrap a few templates, give no free allowance, spend only in one or two countries with strong paying users, and in the ad dashboard compare which creatives get the best click-through and which first-day payments cover acquisition cost. Launch cost is tiny: one website and one 1.0.0 version is enough, and within days you know whether a creative angle works.

The low barrier also means many players, often running the "accounts die" playbook: attach any LLC as the developer name, push the creatives a bit far, and when complaints or delistings come, switch to the next account. Tomi's reviews of double charges, missing points, videos without sound and no support link in the app are not surprises in this playbook.

If it were you

There's something to learn, but don't copy it blindly. This slot is already crowded with identical shells and teams launching accounts in bulk, and Tomi is itself a small copycat. If you follow it, you're competing on creative volume and ad budget, and the edge goes to whoever is willing to swap accounts when they get shut down.

Who it fits: people already running paid acquisition, with a developer account and creative team on hand, who want to test a new theme or country cheaply. Treat it as a test, not a product.

Who it doesn't fit: anyone building a long-term product or hoping to grow through word of mouth and content (this kind has no organic buzz, and users pay once and leave); anyone unwilling to touch borderline creatives or unable to handle refund disputes and delisting risk; indie devs with no paid-ads experience.

How to do it: if you go in anyway, don't make "another fantasy template." Pick a concrete scenario, such as holiday greeting videos, pets, or old photos of deceased relatives. Localize the screenshots for the target country (Tomi didn't even translate its screenshots). Offer a cheap first purchase instead of zero free, and put "failed credits are refunded" and "in-app support" on the store page to cut bad reviews and refunds.

How to get users: start with a small budget on Meta and Google app ads, using "original photo → finished video" comparisons as creatives. Run one country only, check each creative for first-week payback, and stop if it fails. Before anything, scan the top 150 entertainment and photo category top grossing charts in your target country and count how many copycats are already there.

When to move: apps like this can launch in two or three weeks and have no first-mover advantage. If you try, get your first data before the 2026 December holiday creative season, and stop if the numbers don't come.

What would mean we got it wrong: ① If Tomi, Afterglow and this batch of shells are still holding Korea and Japan top grossing a month or two from now, and some team ships a same-type app rated above 4 stars, there's room to build a better version, and it's worth looking again. ② Conversely, if Google Play starts delisting these apps in bulk or tightens its AI-generated content policy, don't bother testing.

---

Data and sources: Google Play store pages (Korea and US versions, 11-06, descriptions, screenshots, reviews); the tomi.buzz website and privacy policy (11-06); Google Play install counts and the Korea and Japan charts (10-19 to 11-06, daily); public content on YouTube, TikTok and X. Acquisition channels, creatives and revenue are inferred, since no external data is public.