AI character chat where "90% of it is ads": a Hanoi app factory turned Character.AI into a watch-ads-for-chat version

Amora is an AI roleplay chat app made by Starvia, a small company in Hanoi, Vietnam. It launched on Google Play in Aug 11, reached 6k installs by Nov 826, and added 7k more in the last 168 days. The product itself is no different from Character.AI or Talkie. What sets it apart is how it makes money: there's no in-app purchase on Play at all. Chatting costs "hearts," and hearts are earned by watching ads. Below: how it took off, where the money comes from, who it's competing with, and whether this playbook is worth copying.

A one-star review: uninstalled before the chat even got going

Sep 14, 2026, a user named Cosmic Monkey left a review on Google Play for Amora: "This is the app with the most ads I've ever seen, 90% of it is ads. Ad after ad after ad. I didn't even get to chat with a character before I closed it and uninstalled." 36 people marked it helpful, and as of Nov 6 it's still the top review on the store page.

Second place is from the day before: "Sometimes it just doesn't reply, and it still takes a heart after it times out. Even basic chat makes you watch a pile of ads to earn hearts. If it's ads in exchange for premium chat or images, fine. But making me watch ads for the most basic chat is not okay." Third is from Sep 30: "Opens and throws an ad at you. Most of the screen is an ad that keeps playing, and you have to keep hitting minimize."

On the same store page, Amora averages 4.5 stars across 11.5k ratings. The bad reviews sit at the very top, yet installs go up every day.

What it is

The store description is the standard pitch for this kind of app: chat with AI characters that feel alive, and the character remembers what you've talked about; characters range from heroes and lovers to villains and sidekicks, across fantasy, romance, adventure, anime, and sci-fi; you can build your own character without code, setting personality, backstory, and voice; text or voice chat. The homepage of the website, amoratalk.com, says "Chat with AI characters. Shape your story." and the sample dialogue line is "The city's still awake. Want to take the long way?"

I couldn't find a single feature that Character.AI, PolyBuzz, or Talkie doesn't have. The one real difference: the Play listing says "Contains ads" with no in-app purchases. Chats cost hearts, and hearts come mainly from watching rewarded videos. The website says "In-app purchases available," which might refer to the iOS version or something planned, not what's on Play right now (unverified).

Two more small details: the Play content rating is "Everyone," but the website FAQ says it's "for adults"; and the data safety section says "No data collected."

How fast it grew

• Installs: Launched Aug 11. The first three weeks barely moved, and by Sep 2 the total was only about 4,300. It started picking up after the version 1.0.3 release in Sep 16: Sep 25 64k, Oct 4 180k, Oct 21 570k, Nov 6 826k (Google Play installs; October and earlier are third-party estimates). The last week averaged over 20k a day.
• Ratings: 11.5k in total, adding a hundred or two a day, roughly one rating per 70 installs, which is high for the category. The top-liked reviews are all complaining about ads, yet the average is 4.5. That suggests the app prompted for a rating inside the app (inferred).
• Charts: Only ever charted on the Entertainment Top Free chart. Saudi Arabia dropped from 90 to 152, India from 101 to 177, and the UAE only sat around #190 for two days in October, on 29 and 30 (Google Play charts, 10-29 to 11-06). It has never charted on any Top Grossing chart.

Over 20k installs a day yet ranking only around #100+ on India's Entertainment Top Free chart suggests the volume is spread across many countries, most likely cheaper regions like South Asia, the Middle East, Southeast Asia, and Latin America (inferred).

Where the traffic comes from

Paid acquisition, very likely, but not directly observed. Install growth isn't a smooth line. It goes flat, jumps, then flattens again: Oct 4 added over 70k in one day, October 8 and 9 each brought in 110k over two days, and Oct 21 brought in 77k. That shape looks like batches of paid campaigns. But store install counts only update every few days, so some of the steps come from how the numbers are counted and can't all be credited to ad spend. The harder evidence is from the same company: its other AI chat app, "Zenda AI", runs two Facebook pages, and in Nov 6 over 400 video ads were live in total, the earliest starting Oct 15, all linking straight to Google Play. Amora's own ads weren't found. By scale, it looks more like Google's app promotion ads or rewarded traffic from an ad network (inferred).

Bidding on big brands' search terms. The store title is literally "AI Character Roleplay," filling the category keyword. The website FAQ has a question written specifically for it: "Is Amora a replacement for Character.AI, PolyBuzz, or Talkie?" That's aimed at people searching those names.

Not relying on it. No Amora videos turn up on TikTok, nobody discusses it on X or Reddit, and the name "Amora" collides with several other AI companion products. No press coverage, only a few automated listings on APK download sites.

Who is behind it

Starvia is registered at Viet A Building on Duy Tan Street in Hanoi, and its LinkedIn page lists 11–50 employees. But the website and support email on the Play listing belong to another company, Growly, and the two are run by the same people.

Starvia has 10 apps on Google Play with roughly 2.65M installs combined, all launched in the last two or three months: the battery emoji app Battery Emoji Buddies 780k, fitness app Yolo Fitness 670k, AI chat app Zenda 130k, AI photo editor PixGen 110k, plus sleep, beauty, and English-learning apps. Sister company Growly's LunaSpeak Jun launched 27 and has 6.65M installs after just over four months. This is a classic Vietnamese app factory: a batch of apps built from templates, with paid promotion added to whichever one gains traction. Amora is that playbook applied to AI character chat.

Where the money comes from

The math probably works like this (inferred): users in low-cost regions rarely pay but will watch ads. Each chat costs a heart, hearts come from rewarded videos, so every model call is backed by an ad impression. As long as the ad revenue from a new user's first few days covers the acquisition and model costs, the business can work without long-term retention. Those "chatted nothing and uninstalled" reviewers may well have watched several ads before uninstalling.

Revenue isn't public, and from outside you can't tell whether it breaks even. What you can see is the ceiling: it has never charted on a Top Grossing list, its reputation is dragged down by ads, and its chart rank is falling.

Who it's competing with

AI character chat is one of the most crowded categories of 2026, and in the same quarter a batch of small studios launched similar apps (Google Play installs, 11-06):

Nobody has copied Amora specifically, because it's just one of the many copycats in this wave. The most useful comparison is Afizzy: launched the same week at a similar scale, one used in-app purchases and reached the Top Grossing chart in France, while the other was pure ads and only hit #100+ on the Free charts in India and Saudi Arabia.

If it were you

Verdict: There's a tactic to learn here, but copying it wholesale isn't a good deal. "Watch ads to earn chat turns" can really drive volume for AI chat in cheap regions. But Amora's volume came from a factory's existing acquisition skills and from ten apps rotating through testing, not from the product itself. The product has no real barrier, and its reputation and chart position are already declining.

Who it suits: Tool app teams that already have experience with ad monetization and paid acquisition (people who've shipped pure-ad products like cleaners, battery apps, or wallpaper apps) and want to apply the same playbook to AI. Also teams with an existing AI chat product that want to launch a free sister app for India, the Middle East, or Southeast Asia.

Who it doesn't suit: Indie developers who can't run paid campaigns and are counting on organic traffic. Amora shows no trace of organic spread; every install was bought. It also doesn't suit anyone aiming to win on product experience, since this approach trades experience for ad revenue, and the top review is "90% of it is ads."

How to do it: Don't copy "watch ads even for basic chat." The reviewer already described a better version: basic chat free or with a generous daily allowance, with ads buying "extras" like longer replies, image generation, voice, and speed-ups. Use the cheapest model tier so each reply costs a fraction of the ad revenue. Offer an ad-free monthly pass for users who will pay; Afizzy shows that some of the same audience will. For languages, pick one nobody is seriously localizing: Hindi, Arabic, Indonesian, or Portuguese.

How to get users: Claim the category keywords and "Character.AI / Talkie alternative" search terms through the store title and website, which costs nothing. Everything else has to be bought. Test a few cheap countries with a small budget on Google app promotion ads, and work out whether the first 7 days of ad revenue per user cover install plus model costs. Scale only if they do.

Timing: A new batch of AI chat apps launches every quarter, so installation prices in cheap regions keep rising. Google Play is also steadily tightening its rules against experiences where ads block core features, so the window for pure-ad models won't stay open long. If you're going to do it, you need a payback model within three months.

What would show I got this wrong: ① Amora's installs keep climbing steadily even without paid spend, and its chart rank in India or Saudi Arabia recovers, which would mean this kind of product can grow on word of mouth in cheap regions and the barrier is lower than I've assessed. ② Starvia adds in-app purchases to Amora and makes the Top Grossing chart, which would mean pure ads are just a starting lever for user acquisition and the real money is in in-app purchases. ③ Conversely, if Amora drops off every chart in December and Starvia stops updating it, that would mean pure-ad AI chat doesn't pay back, and even the factory only runs one round.

Data and sources

• Google Play store page for ai.starvia.friendi (11-06 read): Contains ads, no in-app purchases, 500K+ downloads, 4.5 stars, content rating, data safety, top three reviews (9-13, 9-14, 9-30), developer address.
• Google Play installs, rating count, and Entertainment Top Free chart rank (10-29 to 11-06); daily installs through October and earlier are third-party estimates. Install counts for apps from the same developer and sister company Growly (11-06).
• Meta Ad Library: live ads from Zenda AI's two Facebook pages (11-06). Amora's own ads weren't found, which doesn't mean it isn't advertising.
• Website amoratalk.com (11-06 read); starvia.ai, growly.life, and Starvia's company page on LinkedIn.
• Searches for "amora ai" on TikTok, X, and Reddit: no relevant content.
• Where it says "inferred" or "unverified," it means inference or unchecked.