"The pop-up ads you can't close": how a Vietnamese app factory used blood sugar anxiety to buy 130k installs

Kova is an Android-only app: photograph a meal, get calorie and nutrition estimates, then log a blood sugar reading by hand to see both on one chart. It launched in Jul 22 and reached 130k installs by Nov 6. It's not an indie developer's lucky hit. It's a subsidiary of Apero, a Vietnamese app factory, bought country by country with Facebook ads and a scary landing page. Below: who made it, where the installs and money come from, who it's competing with, and what you can take from it.

An ad you can't close

2026 年 9 月 10 日,一个叫 Ken Webb 的用户在 Google Play 给 Kova 留了一条评论:「Couldn't close a pop up Ad. I didn't get started on the App.」(关不掉弹窗广告,我连 App 都还没开始用。)9 月 25 日又有一条,13 个人点了「有用」:「Scam. Every single click from the get go an unskipable 5 second ad.」(骗局,从一开始每点一下就是一条跳不过的 5 秒广告。)

130k installs, and only 3 written reviews. These are two of them. The third is from Oct 2: "Overall it works well, but some specific foods are hard to log."

In the same period, around Sep 15, a Facebook page called "Kova - AI Scan Foods" started posting image cards: "Know what's on your plate before your next bite. Try the free food scanner now," with a "Scan Food Free" button that opened Google Play. Three days later, on Sep 18, a landing page for Kova went up on the developer's website. Together these three things are the whole story of the app: ads bring people in, they see more ads first, then the paywall.

Who made it

The developer on the store is Astronex AI. Its website footer reads "Astronex by Apero Group," the company is Vietnam's ASTRONEX JOINT STOCK COMPANY, the hotline is a +84 number, and the contact email is [email protected], where the username is "ASO." The CEO's LinkedIn title is "CEO at Astronex (Apero Group)."

Apero is an app factory in Hanoi that says its apps and games have 1B downloads in total. Its PDF reader has over 9,000 million installs on Play. Astronex is its new line this year. Since June 2026, the account has launched an AI English-speaking trainer Elyra (1.78M installs), AI photo retouching Lura (500k), Kova (130k), and AI hairstyle Glovia (15k). The categories have nothing in common (Google Play install counts, 11-06).

The reading is straightforward: run several small AI categories at once, and add budget to whichever one works. Kova is one small note in that set.

What it is

The first line of the store page: "Eating well and managing blood sugar are two goals that belong together." Two main features: photograph food and have AI log calories, protein, carbs and fat; manually log blood sugar before and after meals, on waking, and before and after exercise, then view trends. There is also a home-built "Kova Food Impact Score" that rates each meal from 1 to 10.

It doesn't connect to glucose meters or CGM (continuous glucose monitor) patches. All blood sugar is entered by hand. The store lists it as containing ads, with in-app purchases from $19.99 to $149.99 (Google Play, 11-06). So it collects money from both ads and subscriptions, a standard Apero move.

Where the traffic comes from

Facebook app install ads. The posts on that page are standard install ad cards. One caption reads "Foodient knows YOU": Foodient is the name of another product, which shows the creatives were mass-produced and one was missed. How many ads ran and in which countries isn't public.

A fear-marketing landing page. The headline on astronex.ai/kova reads "Non-invasive blood sugar insight through the food you eat." Below it comes a price anchor: traditional medical care with CGM costs "over $3,500 a year, even with insurance," versus "no hardware, no hidden fees, save up to 85% a year." Further down are four images: vision damage, stroke risk, nerve damage, and heart complications, under the heading "Undetected blood sugar spikes do their damage slowly." This page isn't written for search engines. It's written for the people the ads send in.

Country-by-country launches (inferred). Kova never charted in overall or Top Grossing lists. It only appeared in the free charts for Health & Fitness, ranked between 10-29 and 11-06. Canada climbed from #139 to #42, Spain went from 118 to 52, and the Philippines from 147 to 73. The UK, France, Mexico and the UAE only entered the charts after 11-03 (Google Play category charts). The rankings were pushed up one country at a time, not through word of mouth.

Store search. The title is "AI Food & Sugar Tracker," packing the big keywords food, sugar, calorie and glucose into the title and description. How much of the install volume comes from organic search isn't known.

Almost no presence on social platforms. Searching "Kova" on TikTok returns only unrelated products with the same name, and there's no discussion on X or Reddit.

Install trend (estimated by third-party data sites): 7-29 started at only 543 → about 10k by 8-17 → about 45k by 9-17 → about 109k by 10-21 → 11-06 131k. In mid-to-late October it added a few thousand a day at most, and in November it fell back to one or two thousand a day. In the last 7 days it added 12,360 (Google Play, 11-06). Whether they're cutting budget or switching countries can't be seen from outside.

Where the money comes from, and whether it sticks

Money comes in two stages (inferred): first, app-open and per-tap interstitial ads, the same ad Ken Webb couldn't close; then subscriptions at $19.99–$149.99, likely weekly and annual with a trial, following the usual pattern for apps like this. Android health users in countries like the Philippines and Argentina are cheap to acquire, and ad revenue can cover part of the install cost. Canada, the UK, Spain, Italy and France are where subscriptions have to pay back the spend. It has never charted in the Top Grossing list of any country, so subscription revenue is probably small (inferred).

Retention doesn't show any good news. Ratings: 260, average 3.44. The breakdown is 1-star 86, 2-star 0, 3-star 0, 4-star 57, and 5-star 115 (Google Play, 11-06). The scores bunch up at both ends with nothing in the middle, which looks like the in-app rating prompt only reached some users, plus a batch of people angered by the ads giving 1 stars (inferred, not verified). 260 ÷ 130K, a rating rate of 0.2%, and 3 written reviews. That's what low-intent users bought through paid acquisition look like.

Who it's competing with

"Photo food logging + glucose logging" is not new. Technically Kova is weaker than SNAQ. Its real "difference" is in the ad pitch: framing CGMs as the expensive, cumbersome alternative, then scaring people with complications. On Google Play, several factory apps that bought ads on the word "blood sugar" already have tens of millions of installs. Kova is a small player in this category.

If it were you

C. You can learn a paid-acquisition playbook from it, but it isn't worth copying wholesale. Kova has not yet proven it makes money (it has never charted in the Top Grossing list, and its ratings and reviews are flashing warnings). It got off the ground because of Apero's existing ad team, monetization framework and test budget. And the two categories it has squeezed into, Android blood sugar logging and AI calorie tracking by photo, already have tens-of-millions-install factory apps and leading products standing in front of it.

Who it suits: small teams that already run Meta app install ads on Android, can use ad monetization to offset acquisition costs, and are testing several categories at once. For that kind of team, Kova is a category combination worth trying in a portfolio.

Who it doesn't suit: indie developers with no ad experience or budget who are counting on organic installs; and people who want to build a product that really helps diabetics. That path's competitors are SNAQ, mySugr and the CGM makers themselves, and the contest there is device integration and accuracy, not ad spend.

How to do it: three moves are worth taking. First, category stacking: combine two keywords already proven to drive paid installs (AI calorie tracking × blood sugar) into one new selling point, which sidesteps head-on competition for a single big keyword in the store. Second, a price anchor that substitutes for expensive hardware: comparisons like "CGM costs $3,500 a year, we save you 85%" work well on health paywalls and landing pages; the same framing could be applied to blood pressure monitors, sleep trackers and hearing aids (inferred). Third, a factory portfolio: run three or four small AI categories under one account, share one acquisition and monetization framework, and drop whichever fails to scale within two months. If you want to build a better version than Kova, the most obvious gaps are integrating Health Connect and glucose meters, and cutting ads from "per tap" down to something workable.

How to acquire users: Facebook app install ads, starting with cheap countries like the Philippines and Argentina to collect chart positions and ad revenue, then pushing creatives that work into subscription-rich countries like Canada, the UK and Spain. You can borrow the landing page structure, but complication images and hints like "noninvasive glucose checking" that imply medical efficacy risk takedowns and account bans under Google Play's health policy and Meta's ad policy. Don't copy them.

Deadline: there's no clear deadline. This playbook has been running in Play's health category for several years, so the window won't suddenly close. What will close is platform policy: if Google or Meta tightens rules on health efficacy claims one day, this whole category of creatives fails with it.

What would show this read is wrong: ① if Kova charts in the top 100 of a health category in any country between 11 and December, that suggests the subscription actually pays back, and the grade should go up; ② if its written reviews start showing retention signals like "I use it every day" or "helped me see which foods spike my blood sugar," that means demand is more real than the ad spend suggests; ③ conversely, if it stops updating before December and daily installs fall to a few hundred, Apero has abandoned the bet and there's no playbook worth studying.

Data and sources: Google Play store page, install counts, rating distribution, category charts (10-29 to 11-06) and three public reviews; a third-party data site's estimate of installs since July; astronex.ai's /kova, /about-us and /privacy-policy pages; public posts on the Facebook page "Kova - AI Scan Foods"; the Astronex CEO's LinkedIn profile; apero.vn. Competitor scale is Google Play installs.