A 21-rating tool hits #32 on the Top Grossing chart: how an AI product-video app collects money by charging per second

In Nov 4, a user named "槐不语" left BeeShow a one-star review on the App Store: "It can only generate 15 seconds at most, and it costs 5 coins per second. 15 seconds is 75 coins, and even after buying a membership you only get a few 15-second videos." The review is harsh, but it accidentally explains exactly how the app makes money. BeeShow launched on the China App Store in Sep 20. It does one thing: you upload a product photo and it generates a short sales video with AI. By 11-06 it had only 21 ratings in total, yet it ranked #32 on the China Top Grossing chart for Utilities, ahead of several competitors with dozens to hundreds of times more ratings. Below: what it sells, why people pay, where its users come from, who it's competing with, and whether you can still get in.

What it is

An iPhone app whose title is basically a search query: "BeeShow - One-click original e-commerce sales videos." The store page lists six features: one-click sales video from a product photo, replicating the structure of viral videos, generating voiceover from a single selling point, placing the product in a short storyline, "authentic" UGC-style seeding, and multi-style product display. Output is vertical short video for sellers posting on Douyin, WeChat Channels, and Xiaohongshu.

Monetization has two layers. The first is membership: weekly, monthly, quarterly, and annual auto-renewing tiers. The second is that generating video costs extra coins, 5 coins per second, up to 15 seconds per clip, which is 75 coins per clip. The coins included with membership only cover a few clips.

How 21 ratings got it to #32 on Top Grossing

First, its trend over the past ten days on the China App Store, Utilities category:

Top Grossing held steady at 27 to 52, while Top Free bounced between #75 and #201, and on one day it fell off the chart entirely. Few downloads, but the people who pay spend a lot. Compare it with its peers on the same day (11-06): Quick Edit Pro has 12,739 ratings and only ranks #174 to #189 on Top Grossing; Hit Studio has 2,166 ratings and peaked at #58; Instant Cut AI has 706 ratings and peaked at #159. BeeShow's rating count is a fraction of theirs, yet its revenue rank is higher than all of them. 11-06 On Top Grossing it sits right next to long-running paid tools like voice-to-text transcription, wallpapers, and Sogou Input.

The cause is the "double charge" from the one-star review: a subscription, plus coins charged per second on top. Sellers and newcomers willing to pay for sales results may buy a membership and then top up coins several times, so the average spend per user is far above a typical utility app. Exactly how much revenue this rank represents can't be calculated from the outside.

The cost is reputation. From 10-27 to 10-30, a string of 5 one-star reviews came in, all about money: "Bought a membership and it still charges separately to generate videos," "You have to pay the moment it generates, and the membership costs more after that," "No need to sign up, but spam floods in," "This boss has a small outlook." Starting 11-01, the tone suddenly shifted, and 5 five-star reviews arrived in a row, with polished wording like: "Plain product intros can feel stiff, while the storyline sales feature weaves the product into everyday scenes... easier for users to keep watching than just reading out specs." Whether these positive reviews were solicited or astroturfed, the wording and timing look similar, but that's only an inference. 11-06 The store average is 3.1,23 ratings.

Who makes it: a Jiangsu team running a matrix of utility apps

商店页上的开发者是一家香港公司,RENREN MOBILE GAMES NETWORK TECHNOLOGY CO., LIMITED(人人手游網路科技有限公司),2024 年 1 月注册,账号下只有这一个 App,开发者编号和 App 编号几乎挨着,看得出账号是专门为它开的。名字容易让人联想到老牌的人人游戏,但那家的主体是上海人人游戏科技发展有限公司,两者看不出有关系。

The one actually running it is Nantong Tuiliang Network Technology Co., Ltd. in Jiangsu. BeeShow's Douyin official account is certified under them, and they're the developer of the Android version (on Tencent's App Treasure). The app's privacy policy is hosted on a backend called xiaofengle, whose homepage reads "Multi-App Foundational Capability Management System: unified login, payments, memberships, push, files and data services." The same backend also runs an AI dubbing app, "Xiao Peiyin." Nantong Tuiliang's own Apple developer account also has "Mi Xiaoyu - the chat keyboard that gets you" (launched 2024-06, 11.5k ratings in the China store, best rank 73 on the category grossing chart).

In short, this is a team with an existing backend that knows how to build paid utilities. For them, launching another app is a reskin plus a new video generation capability. Publishing under a fresh Hong Kong account also means problems with one app won't spill over to the others (inference).

Where users come from

Likely paid acquisition on Douyin (inferred, not directly seen). BeeShow's Douyin account is IP-located in Jiangsu, with 1,681 followers. Its public profile shows only 7 videos, with 2 to 22 likes each, adding up to just over fifty, yet the account's total likes read 11,324. The most plausible explanation: most of the likes sit on ad creatives that aren't on the public profile. Ads run through Oceanengine, Qianchuan, and DOU+ often look like this (inference). The company is called 推量. The public videos read like ad copy. The one from 11-03 says: "Tired of shooting real product videos? Try this: one image can make a viral short-drama product video... no filming needed, one image is enough." The one from 10-12 says: "Everyone keeps asking how to use it, here's the AI product video tutorial!" (Douyin, viewed 11-06)

Store search (confirmed). The title crams in "sales video, original, one-click generation." The category is "Utilities"; similar apps often sit under "Photo & Video," but the Utilities Top Grossing bar is lower, so it's easier to chart. In the App Store "Customers also bought" section it appears alongside 素剪AI, 天天爆款, 瞬剪AI, 探思AI, and 天天出单, meaning the same group of beginner AI sellers tries several of these apps back to back.

Download-site coverage (confirmed). Huoniao Game, Yuanpi App Store, Yingyong Kong, Youba Le and Kefan Download Site all host its Android download pages, the standard SEO play for domestic tool apps. But the Android side isn't really doing anything: Tencent App Treasure shows only 21 downloads (11-06).

What I couldn't find: no press coverage, and no discussion on X, Reddit, or Xiaohongshu. Revenue is not disclosed.

Why now

In 2026, "AI sales videos" were hot on Douyin. The most widely circulated case in Zhihu and operator circles is 雪梨爆爆: 110k followers, sales of 10M to 25M in three months, and over 5200 videos posted, almost all AI-generated women's apparel sales clips. A wave of newcomers wanted to copy it, and "one photo to sales video" tools were built for exactly them. The underlying video generation models also got cheaper this year, so a small team can wrap a shell around them and ship.

But BeeShow wasn't the only one riding the wave. On Douyin, the tutorials teaching "AI storyline sales" mostly use ByteDance's own Xiaoyunque. A account called "木琴" posted two clips in September, each with over 1,400 likes, dozens of times more than any video on BeeShow's own account. What BeeShow sells is really big-tech capability packaged as "sales-specific templates," plus convenience.

Who it's competing with

Overseas there's CreatOK (aimed at TikTok Shop sellers, pitched as "one-click viral copy"), Creatify, Arcads, MakeUGC, Oumomo, and others. Most are web-based with USD subscriptions and do the same thing.

Nearly every one of BeeShow's six features has a counterpart among competitors. "Generate directly from a product photo, add storyline and seeding" is the newer generation compared to the "mashup and dedup" camp, but the whole segment is moving this way. Where it truly differs is at the business level: heavier charging, clever category choice, and paid acquisition skill. It got where it is not because its product is better.

If it were you

Verdict: Don't copy this app as-is. Before it launched, China already had at least 8 similar "AI sales video" tools, and ByteDance's own tools sit above them. What's worth learning is how it collects money: subscription plus pay-per-use coins, sold to people willing to pay for sales results. A few dozen ratings can be enough to chart on Top Grossing.

Who it suits: Small teams that already have login, payment, and membership infrastructure, that run ads on Douyin or TikTok, and can tell when they've paid back their ad spend; or people who already have a network of sellers or sales creators.

Who it doesn't suit: Solo developers who can code but have never run paid ads. This category has no technical moat, and the product can't be differentiated on features, so the game is creatives and bids. It's also not for anyone who cares about reputation or wants to build a long-term brand. This charging model produces one-star reviews.

How to do it: Two paths. One is to change the audience: take the same "one photo to sales video" model to TikTok Shop sellers in the US store and Southeast Asia, localize voiceover and subtitles into local languages, and charge subscription plus per-second fees. Note that CreatOK and Creatify already work the US store, and I haven't checked how crowded the Southeast Asian language segment is. The other is to change verticals: apply its charging structure to another audience that spends money to make money, such as property listing videos for real estate agents, group-buy videos for local restaurants, or main-image videos for cross-border sellers. Pay-per-use extracts more from heavy users than plain subscriptions. Whichever path you take, show users a complete finished clip after the trial and before coins are charged. Reducing the feeling of being burned ("charged the moment it generates, then the membership costs more") should cut down on bad reviews considerably (inference).

How to get users: Demo the product in your ad creative itself: a white-background product photo on the left, the generated 15-second clip on the right, opening with "Don't want to shoot real footage?" Make the store title a search query, and pick the category with the lowest Top Grossing bar. On Douyin or TikTok, test creatives on a small budget first, and only scale spend once you've calculated how many coins a paying user actually burns.

When it's already too late: The window in China is basically closed. The overseas TikTok Shop route, from the Black Friday to Christmas stretch (only three weeks away now), is too tight to build from zero. At the latest, launch by 2027 March and test creatives on a small budget so you're ready for the mid-year sales push.

Signs you've misread it: ① BeeShow drops out of the top 100 Utilities Top Grossing within a month, meaning refunds and churn from per-second coin charges are eating the revenue, and this pricing won't last. ② ByteDance builds "one-click sales video from a product photo" directly into Xiaoyunque or CapCut and makes it free, which takes away the entire category's shovels. ③ Conversely, if a month from now it still has only dozens of ratings and stays within #50 on Top Grossing, the "few users, high ticket" model is more durable than it looks, and it's worth copying seriously into other verticals.