Searching "Lovart" on the App Store, page one has no Lovart: how a borrowed-name app squeezed onto the China design bestseller chart

Lovart is a design agent from LiblibAI. It was hot in China's design scene for a full year, but it only has a web version, and there is no official Lovart app on the App Store. A developer account with a Romanian name first listed the app under a different name, then renamed it "Lovart AI" and caught people searching for the brand. Worldwide it has only 51 ratings, yet it reached #55 on China's graphic design bestseller chart. This piece covers how it got onto the chart, where the money comes from, how many others are doing the same thing, and whether you should.

Nov 6, searching "Lovart"

In 2026 Nov 6, searching "Lovart" on the China App Store, the first result is "Loveart Poster Design," published by a Shenzhen company in January, with 84 ratings. The second is "Lovart AI: Poster & Brand" by developer Teodora Ilie, published in Sep 16, with 43 ratings in China. Below that are Kling, Doubao, Coze, and an app called "Seedreem 5.0." LiblibAI doesn't appear at all, and nothing official from Lovart shows up.

Searching the same term in the US store gives much the same results: the first is Teodora Ilie's app, the second "Lovart AI - Photo Generator," published in May by a company called Ignite Classes, and the fifth, "Gimage 2.5," is from the same developer as the Seedreem app in China (App Store search, 11-06).

The real Lovart launched publicly in July 2025, offering an infinite canvas and a design agent. According to reports from 36kr and KuCoin, its ARR passed $80M within 5 months of launch. Its parent company LiblibAI raised about $300M in a B+ round in June 2026 at a $2B valuation. A name this big is an empty lot on the App Store.

What it is

An ordinary "design from one sentence" subscription app: describe what you want, pick a style, and it generates social graphics, posters, ad images, brand kits, and product mockups. The description says it "integrates multiple top image and video generation models into a single subscription," which I infer means it relays calls to third-party model APIs. The package is only 17MB, and the privacy policy is a Google Docs file.

It wasn't originally called Lovart. The store URL path still reads genart-ai-design-agent, and the storefront pages in countries like Mali and Slovenia still show the old name "Genart AI Design Agent – Prompt to Poster, Brand & Ad." It got through review under a generic name, then was renamed to a popular brand. I couldn't pin down when the rename happened. The version stopped at 1.0.2 in Sep 24 and hasn't been updated since.

The same developer account also has "Dance AI - Generator Video" (2026-06), with 5 ratings, and its privacy policy is also a Google Docs file. This is a shell being reused over and over.

Scale

(App Store rating counts and category charts, 10-28 to 11-06)

China accounts for 39/51 of the ratings, and nearly all downloads come from China. From 10-28 to 11-01 China had no new ratings at all. Starting 11-02, it adds 1, 4, 4, 0, and 3 per day. I don't know why it suddenly sped up. It could be that the real Lovart got another round of discussion in China, or it could be fake ratings.

The rankings in other storefronts look impressive, but they're cheap. Argentina 11-05 is still at #135 on the Graphic Design grossing chart, and after gaining 1 ratings in a day it jumped to #29. Uruguay went from #33 to #88, Taiwan from #67 to #92, and it's showing up on category charts in 13 to 17 stores. These are small stores with small categories, where a handful of subscriptions is enough to chart.

The US storefront has 13 ratings, averaging 4.69, and hasn't charted anywhere.

Where the money comes from

The paywall is hard and the prices are high:

(in-app purchase prices on the store page, 11-06)

The China storefront has only two written reviews, both one-star. One from Oct 17: "Don't download this. You get in, you top up, and you can't do anything. It keeps showing generation errors. Pure scam." One from Nov 1: "Often times out... can't make images... really bad." It takes your money the moment you open it, and generation still isn't reliable after you pay.

The math isn't complicated. Traffic from searching the brand name costs nothing, the model cost of generating an image is a few cents to a few tenths of a yuan, and as long as a small share of those searchers buy the ¥58 weekly subscription, that's enough to reach #50–60 on a small category bestseller chart. Ahead of it are Gaoding, Meitu Design Studio, MAKA, and Chuangkit, and beside it are "Gai Zi P Tu" and "Zuotang" (China graphic design bestseller chart, 11-06). Real revenue isn't public. The bestseller rank only shows that someone is paying, and it can't tell whether that's thousands or tens of thousands.

All 5 five-star US ratings fall between Sep 26 and 30. The usernames are all "name plus three digits," and each review praises one selling point: brand kit, product mockups, ordinary language, social graphics, complete set. This is very likely paid or fake reviewing, meant to make the store page look better and convert a few more people.

Beyond that, I found no sign of ad buys, creator outreach, or social media work: no X post mentioning it in nearly 60 days, and the only web presence I could find is its own store page.

Who it's competing with

It is itself a copycat, and not an early one:

The product itself has no scarcity. It only runs for one reason: nobody official has claimed the "Lovart" search term on the App Store.

If it were you

One-line verdict: Don't follow. This is a wrapper riding on a borrowed name, and at least three same-name or near-name copycats are fighting for the same term. The earliest one launched half a year before it and has more volume. Its entire business rests on two things you don't control: the official product not shipping an app, and Apple not acting on trademark complaints. The only idea worth taking: where a brand has blown up on the web but has no official App Store app, the search traffic is sitting empty.

Who it suits: Studios that already mass-publish subscription wrappers, can live with apps being pulled at any time, and treat each app as one-off cash flow. They don't need this article to know how to run it.

Who it doesn't suit: Indie devs trying to build a product that lasts; anyone who cares about developer account safety (trademark complaints plus two App Store Review Guidelines clauses, 4.1 impersonation and 5.2 intellectual property, could get the whole account banned); and anyone who doesn't want to get tangled in "pay and get nothing" bad reviews.

How to do it (if you only want the idea): Don't borrow someone else's name. Borrow the gap. Find a category that is a web hit with no native app from the official owner, build a genuinely usable mobile app under your own brand, name it and write screenshots so the category is clear (e.g. "AI poster design agent"), put category keywords in your metadata rather than someone else's trademark. Then do well where the web version is clumsy on phones: direct photo-library import, one-tap sharing to Xiaohongshu and Moments, pay-per-use instead of an upfront weekly subscription.

How to get users: Same as it does: start with App Store search. Rank on category and long-tail terms. Small sub-category charts in the China store have low thresholds, so even a modest ranking brings some browsing. Then post tutorials on Xiaohongshu like "how to make posters with AI on your phone" and send people to your app.

When it's over: The window isn't in your hands. The day the official Lovart ships a native app, traffic for this term goes straight back to the official side. The day Apple acts on complaints, these borrowed-name apps get pulled together.

What would mean you read it wrong: If by early 2027 Lovart still has no official app, the borrowed-name apps are still around, and China store ratings keep climbing steadily, then the official side doesn't care about mobile and Apple doesn't police it, and this gap can be held longer than expected. If their ratings stall at a few dozen and they drop out of the top 150 on the grossing chart, then the term never had enough search volume to sustain even a wrapper.