A few 80 a day, one sentence 15: Pogee rode paid acquisition and a niche chart to 37k installs and Taiwan's #8 Top Grossing

Pogee is an AI character companion app whose pitch is "the characters keep living their lives when you're not there": they post updates, write diaries, and hide secrets, and the storylines are visual novels with BGM and CG. It began buying Google ads in late September, and in early November it charted at #7 and #8 in Taiwan's Google Play Lifestyle Top Grossing, yet it has only 37k installs on Play in total, with a rating just over 2. Below: what it is, where the ranking came from, how it collects money, and whether the category is worth entering.

A review changed from three stars to one

On 9 2026, Taiwanese user hjenre edited their Google Play review:

You could only get 80 a day, the chat's default replies cost 15, and the conversation ended after a few lines, so you could only continue by paying, though it's freer than any other app... dropped from three stars to one... spent 45 ice cubes to unlock one day's diary and it wouldn't open at all, 45 gone.

18 people marked a review as helpful, the most-agreed review for Pogee in Taiwan. Another user, Oct 15, wrote: "Tried 20 AI chat apps and this is the worst one, no contest. Crashes nonstop... paid for the thought diary and secrets and it just froze."

These two reviews sum up the app. First, it does have things others don't ("freer than any app", with diary and secrets unlocked by paying). Second, the paywall is hard, and ice (its currency) is the gate. Third, lots of bugs. One more thing: under the negative reviews, the official account replied to all of them in one batch on Nov 1, each ending with "If things improve later, feel free to update your review."

What it is: a phone that lets you peek into a character's life

The first line of the store page reads: "When you open Pogee, another world has already been running. Someone just posted something new, someone is living by today's schedule." What users get is a "phone inside the world":

• Chat and voice calls with characters. The end-of-October version added video calls, and action commands make characters move;
• Watch the character's feed, schedule, mood, diary and "inner monologue". Characters message you first when they think of you;
• Visual-novel-style storylines with signature BGM and character and scene CGs. The system gives options, and you can write your own if you don't like them;
• Users can also create their own characters. The new version added a "creator homepage" and "Little World" (management, simulation and puzzle gameplay).

Monetization is currency plus in-app purchases, with Play prices of $0.99–59.99 per item. The Taiwan reviews tally it up this way: daily login gives 80 ice, one preset reply costs 15, and unlocking a day's diary costs 45. The Japan official account's launch headline was "Now get 300 ice".

The publisher is WORLDOS PTE. LTD. in Singapore, registered in Jun 3 2026. The team is most likely in mainland China (inference): Taiwan update notes include simplified-Chinese phrasing like "video call" and "unlock unlimited story play", ad copy says "has voice and memory", and the privacy policy points to Alibaba Cloud Japan. The iOS version launched first in Jul 27, the Play version in Sep 7.

Where the ranking comes from: Google paid acquisition, plus a low-barrier chart

Start with the numbers. Play installs were 11-04 to 36,147,11-06 to 36,773, up 626 on the day. It sat at #8,11-06 on Taiwan's Play Lifestyle top grossing chart for three straight days, from 11-04 to 11-06, and the store page showed #7. On iOS Taiwan, in the same category's top grossing chart, it bounced between #36 and #73, with only 4 ratings on the Taiwan iOS listing.

Now look at its neighbors on the chart. On the same day, #2 on this chart was Xiaomi Home, #3 was TP-Link Tapo, and #9 was a lottery-results app. The Lifestyle top grossing chart is not the main battleground for AI companions; Entertainment is. An app with tens of thousands of installs can crack the top ten here as long as a small group pays. Its most direct Taiwan rival, WHIF (1.6M installs on Play), also uses this category and ranks #3 on the iOS Taiwan Lifestyle top grossing chart.

Finally, where the traffic comes from. In Google's Ads Transparency Center, advertiser WORLDOS PTE. LTD. has 21 creatives in total, paid through a company called Sinointeractive Digital Marketing, with landing page pogee.ai:

• Starting Sep 23, app-install video ads ran in South Korea, Japan, and the US and Canada at the same time. Most US, Canada and Korean creatives were paused around Oct 19;
• Taiwan spend started Sep 29, and new creatives were added around Oct 12 and Oct 30, including video, image and search ads, with copy reading "chat, calls and free-input storylines";
• Around Nov 5 there were still 12 creatives running, only in Taiwan and Japan.

This is a standard overseas test: four markets opened at once, then after three to four weeks the money was concentrated in the two that performed best. The ranking climbed from late October into early November, and the timing lines up.

The Japan route: paid note articles plus a Halloween character event

Japan was its second market, and the playbook was different. From Oct 26 to 30, at least 4 review articles appeared on note.com, each opening with "commissioned by Pogee" or tagged 【PR】. Three of them had writers invent a Halloween character (for example, a "vampire who appears only on Halloween night and has lived over a hundred years"), tied to the official "Pogee Halloween 2026 Character Festival" and a feature of works by 21 invited creators. Writer もへ spent 5 days, concluding "it's really fun once you figure out the features on day 3 and the feed starts moving," and topped up 300 yen on their own. Another writer ran out of ice.

The official account has its own "pogee editorial desk" on note, and on pogee.ai it publishes comparison pieces like "Best AI Character Chat Apps". The official TikTok account @pogee_jp had 66–112 views per post. No user-driven discussion turned up on X, Reddit, Dcard or Threads.

Who it's competing with

The AI companion space in Taiwan and Japan is already crowded:

(Install count is Google Play, as of around 11-06)

What's unique to Pogee is only the packaging: it makes the feed, schedule, diary and inner voice the headline features, plus visual-novel-style storylines. Other apps have these features too, they just don't put them up front.

Has it added up?

Not yet. 37k installs, two months of paid acquisition, and a 2-plus rating (global 2.3, Taiwan store page 2.7–2.8). The rank comes from a low-bar category and a hard paywall. What the Taiwan #8 spot actually earns per day isn't visible from outside, could be just a few hundred dollars (est.). Whether it stays on the chart once paid acquisition stops is the main question it has to answer next.

Few people know about it: no media coverage, no public revenue, and what you can find online is almost entirely official content, commissioned pieces and APK sites. That doesn't help followers much, because it hasn't uncovered any new demand. That Taiwan users pay for AI characters has already been proven by a dozen similar apps on the top grossing charts.

If it were you

Verdict: copy its approach, don't copy the product. Character companionship is a red ocean in Taiwan and Japan. Pogee is just a new face that paid its way in, and its reputation is still sliding. What's actually useful is how it used very little money to break into the top ten of the grossing chart.

Who it fits: teams that already have an AI companion or interactive story product, have art and script capacity, and are looking at their next market. For you, Pogee shows a cheap way to get in.

Who it doesn't fit: indie developers trying to build an AI companion app from scratch. Characters, CGs, BGM and storylines are ongoing content costs, model calls are billed per sentence, acquisition means buying installs, and you'd be competing with WHIF, zeta and Talkie for the same people. It's also not for anyone who can't run paid campaigns or is counting on organic traffic: in two months Pogee produced not a single piece of user-driven content.

How to do it (if you already have a product):

• Pick the category. Like Pogee and WHIF, put a small market like Taiwan into a low-barrier category such as Lifestyle. Charting itself is exposure.
• Make "the character's life" the paid feature. Diaries, secrets and inner voice unlocked one by one feel more like content than charging per chat, and users are more willing to pay for them.
• Do what it didn't do well: the free currency given on day one should be enough to finish a full storyline, failed unlocks should refund automatically, and sensitive-word filters shouldn't wrongly block the options the system itself offers. Nearly all the negative Taiwan reviews come down to these three things.

How to acquire users: run Google app-install ads at small budgets across three or four markets at once, then after three to four weeks keep only the one or two with the best payment rates. In Japan, run a round of commissioned note.com articles where writers invent characters, and build them into holiday events (Halloween, Christmas, Valentine's Day). The character link is the acquisition link.

When to move: there's no window here. It's a long-term category where budget and execution decide the outcome. If you want to learn this entry playbook, test your creatives before December's Christmas and January's Lunar New Year, the two periods when character-companion payments peak (inference).

What would mean you read it wrong:

• Pogee stopped paid acquisition, yet a month later it still sits in the Taiwan Lifestyle top 20 with a rating above 3.5. That shows the character-life loop itself keeps people around, so the product format is worth studying seriously.
• Conversely, if it drops out of the top 50 before the end of November: it shows the rankings were all bought, and even the payback of this entry playbook becomes questionable;
• If a second or third new app in Taiwan or Japan makes "characters with their own lives" its main selling point and also runs paid acquisition: the packaging starts getting copied, and its differentiation is gone.