Paused for 17 days, then up 20k in a week after the spend came back: CineDream is a mid-tier shell in a paid-acquisition matrix
What it is A 17+ entertainment app on Google Play where you upload a selfie, pick an effect or outfit, and generate an AI video. Credits are sold per use, $2.99–$99.99.
Why it took off Not because the product took off: the same template sits on a dozen-plus bought developer accounts, paid acquisition buys installs by package and country in cheap markets like Indonesia and India, pausing stops installs and resuming restarts them, and a few high-value first purchases cover the cost.
CHow you could play it Don't build the 13th shell. A team that knows how to run install ads could build one clean single-feature app (bring old photos to life, pet videos), plug into an off-the-shelf video model API, and test first-purchase payback with a small budget in cheap markets.
- Data (updated daily) · first launch 2026-09-03 · 11-07 installs 35,598
Android · Total installs worldwide Data 2026-10-29 → 2026-11-07 (10 days, 10 observations). Google's detail page only refreshes install counts every other day, so a flat day doesn't mean nobody installed.
CineDream – AI Video Magic is a 17+ app on Google Play that turns a single headshot into a cinematic video. It launched in September, has 36k installs, and zero ratings and zero reviews. Its install curve runs for a stretch, stops for half a month, then restarts, like someone turning a tap on and off in the background. It is actually one of at least 12 same-template reskins, and not the largest. What's worth watching is how the hand turning the tap works, not the app.
A curve that stopped for 17 days
CineDream launched on 3 September. From mid-September it gained several hundred installs a day, reaching 4 total by 9,171 October. Then it nearly stopped: from 5 to 21 October, it added fewer than 17 installs over 40 days, one or two per day (AppstoreSpy estimate).
On 22 October the tap reopened. A week later it was at 13,706, and another week later at 35,598. In the 6 days before 7 November it added 21,892 installs (Google Play installs, 11-06). Around the same time it entered the Indonesia and India Entertainment top grossing charts, peaking at #1 in Indonesia on 160 November before sliding to #185 and #190.
An app that grows through word of mouth doesn't look like this: word of mouth doesn't drop to zero for 17 days and then come back exactly on schedule. The most plausible reading of this curve is that the ad budget was paused and then turned back on (inference, high confidence).
What it sells
The store page's five screenshots follow one formula: bottom-left, a plain selfie of a young girl in a white T-shirt; the middle shows the same face, full-body, on a bench under the Eiffel Tower, on a balcony in an old town, on a seaside promenade, in shorts and skirts, with exposed midriffs. The headlines run: "Tap Once Video Comes Alive", "Zero Skills Pro Results Instantly", "Try Trends Without Buying", "Tap Once Outfit Reborn".
Beyond "photo to video", the description has a whole paragraph on AI outfit swapping: "From streetwear to evening gowns... Tap once, your outfit is reborn." Rated Mature 17+, flagged "Diverse Content: Discretion Advised" and "Users Interact". Read the screenshots and the rating together, and the pitch is "put a face into a prettier, sexier video" (inferred; I didn't install the app to check its effects list).
It makes money from per-item in-app purchases, $2.99 to $99.99 each (Google Play store page, 11-06).
One more thing to note: its data safety section says it "may share personal info, photos and videos with third parties", and also "data is not encrypted in transit". Users are uploading their own faces.
Eleven identical siblings
Take the privacy policy URL: truclnerhpsu.play-cine.com/api/quhomadcdscq.html, a random subdomain with a random string. Reverse-searching that pattern on Google Play turns up a whole row of the same kind (Google Play data, 11-06):
| App | Developer account | Launched | Installs | 7 new in the last day |
|---|---|---|---|---|
| HiFrame – Photo to Video AI | GE Technology | 08-12 | 24k | 19k |
| Novora – Image & Video Maker | Unnica Telemetria | 08-19 | 44k | 4k |
| HoneyDrip – AI Video Magic | SupperSoul | 08-20 | 90k | 39k |
| Sweetory – AI Video Generator | Frameverse | 08-20 | 35k | – |
| Filmlet – AI Photo to Video | Lavish Food | 09-02 | 53k | – |
| WowViz – AI Video Creator | WeDo Technologies | 09-02 | 68k | 25k |
| CineDream – AI Video Magic | Zama Zama | 09-03 | 36k | 22k |
| FilmForge: Photo to Video AI | Lavish Food | 09-20 | 74k | 52k |
| LumiCraft – AI Video Experiences | Brady Carter | 09-21 | 32k | – |
| Aurion – AI Video Experiences | Perawat 24 | 10-05 | 38k | 38k |
| VisionWave – Photo to Video AI | silambarasan | 10-14 | 556 | – |
| Woolora: AI Turns Vision Video | GLAD Technologies | 10-16 | 27k | 24k |
What they share: all 17+, all in the Entertainment category, all with in-app purchases of $2.99–$99.99, all named along the lines of "X – AI Video Magic / Photo to Video AI / AI Video Experiences". Their privacy policies are either "random subdomain/api/random string.html" or, in newer builds since late September, a random CloudFront address plus a 8-character random filename. Install counts for 10-31, 11-03 and 11-05 jumped on the same day.
CineDream's developer account is "Zama Zama", its registered legal name is a Latin American personal name, and the contact email is a personal Gmail. The other accounts' legal names span Indonesia, Pakistan, Nigeria, Latin America and the US, and one account called "Lavish Food", a food name, carries two video apps. Mismatched account names, legal names and emails look like purchased or rented developer accounts (inferred, high confidence). The twelve apps add up to roughly 520k installs, most likely from one team or a paid acquisition studio (inferred).
Where the traffic comes from
Bought. 36k installs and not a single rating. Its privacy policy lists AppsFlyer as an SDK, plus Google's install referrer, ad IDs and Play billing. That combination exists to work out how much revenue each ad dollar brings back. Off-store there is no organic buzz: no TikTok videos for it, zero posts on X in the past 60 days, and only auto-generated APK download-site pages on the web. Whether the spend went to Google, Meta or ad networks can't be seen from outside.
Split by country and by app. CineDream charted only in Indonesia and India, while its sibling apps charted in Italy, Mexico, Malaysia and elsewhere. Each app tests one set of countries: if it runs, spend goes up, and if it doesn't, it gets shut off. CineDream's 17-day gap is what "shut off" looks like.
The name didn't help. "CineDream" happens to share its name with a well-known color-grading plugin (cinedream.io, which video editors know), and searching the term mostly returns the plugin. It doesn't look deliberate, and it did little.
Where the money comes from
It sells credit packs, with the top tier at $99.99. Installs in Indonesia and India are cheap, so a few curious people paying once can cover part of the cost. Charting at 160–190 in Entertainment Top Grossing shows real revenue, but not much of it; in this category, ranking beyond 150 is easy to reach. Revenue per app is unclear. That the operation can launch a new app every week or two suggests the whole matrix probably pays for itself (inferred).
Who's competing
Outside the matrix, companies doing the same thing are two to three orders of magnitude bigger (Google Play data, 11-06): PixVerse at roughly 80M installs, Vidix at 30.2M, Nexus at 12.4M (+1.85M in 7 days), Phant AI at 12M, plus the model makers' own apps (Kling, Hailuo, Pika, Vidu), and Sora, Veo, and CapCut's built-in "photo animation". CineDream has no model of its own. Its privacy policy even contains boilerplate about "fine-tuning Stable Diffusion to generate avatars", which doesn't match a video app.
The product has no differentiation. Its difference is entirely in how it operates: splitting ad spend and ban risk into a dozen small pieces.
If it were you
The verdict: copying the 13th shell isn't worth it. The playbook, one template, many accounts, test countries and track returns, is worth learning, but at least one studio has already saturated it, and a solo developer's single account can't compete.
Who it suits: small teams that already run install ads, hold several developer accounts, and can absorb batch bans.
Who it doesn't suit: indie developers who want to survive on product quality and organic traffic. Here users install, pay once and leave, and zero reviews is the evidence.
If you really want to build it: skip the borderline outfit-swap angle. Pick one clean use case that people will pay for (say, animating old family photos, or turning pet photos into video), plug in an existing video model API, state your privacy practices honestly, sell per item, and price the first purchase to cover acquisition cost.
How to get users: vertical before-and-after short-video creative. Test with small budgets in cheap-install markets like Indonesia, India and the Philippines, measuring first-purchase return, and add countries only when a test pays back.
When to move: once video model API prices drop, this type of shell multiplies. Google Play could crack down on linked accounts and 17+ borderline content at any time, so testing sooner is better, but don't put everything on one account.
What would prove this read wrong: if these twelve apps get pulled in batches over the next few months, or no new ones appear, the math has stopped working. Conversely, if one of them breaks into the top 50 of Entertainment Top Grossing, these shells earn more than they look like they do, and it's worth looking again.
Research notes: competitor table, acquisition evidence, full research report
- Competitors
Name Where Launched Scale How it differs from this one PixVerse Android 2024-09 About 80M installs, 7 days +520k (googleplay library) Own model and brand, global ads; CineDream is just a template reskin Nexus - AI Video Generator Android 2026-03 12.4M installs, +1.85M in 7 days A photo-to-video ad tool too, but with a big-spend flagship; CineDream runs many small apps under multiple accounts Vidix-AI Photo Video Generator Android 2025-11 30.2M installs Large single app, all ages; CineDream is 17+ borderline, charts in only two countries Phant AI Android 2025-06 12M installs, #9 in Nigeria Entertainment Top Grossing Similar effect templates, but already an established brand HoneyDrip – AI Video Magic and other same-template clones Android 2026-08 to 10 0.05–90k installs each, about 520k combined Almost no difference: swapped name, icon and account Kling / Hailuo / Pika / Vidu / Sora / Veo Big-company features 2024–2025 Tens of millions of users Run by the model maker, stronger quality and brand; CineDream only does templated packaging and ad buying - Where users come from
- Inferred Paid acquisition (channel not verified): privacy policy lists AppsFlyer, Google Install Referrer, Google Ads ID/AppSet ID; 35k installs, 0 rating, 0 reviews, no organic content off-store; the ad library doesn't cover this app
- Confirmed Clone matrix / reskinned mass distribution: 12 apps in the googleplay library share Mature 17+, Entertainment, and in-app purchases of $2.99–$99.99; privacy policies are random subdomain /api/random string.html or random CloudFront pages; developers' legal names differ and are mostly personal Gmail; install counts jumped the same day
- Inferred Name piggybacking: CineDream shares its name with the Color Grading Central color plugin, and searches return only the plugin; no evidence of deliberate piggybacking, and the effect is small
- Inferred TikTok / X / media: TikHub and x-radar searches find no content for this app; the web only has auto-generated APK download pages
Why CineDream – AI Video Magic (com.generator.cinedream) was able to run
Platform: Google Play | Developer: Zama Zama (legal name ERIC LOPEZ, email [email protected]) | Listed 2026-09-03 | Entertainment | Mature 17+ | In-app purchases $2.99–$99.99 | Research date 2026-11-06. Data from the googleplay library (app / app_day / chart_rank, from 2026-10-29), the Play store page, the privacy policy page, APKPure, TikHub, x-radar, and Jina search.
Bottom line
CineDream isn't an independent product that broke out. It's one shell in a templated reskin matrix. Between 2026-08 and October, at least 12 nearly identical "photo-to-video + AI outfit swap" apps were listed under 12 different "developer accounts", all Mature 17+, all with in-app purchases of $2.99–$99.99, and all with privacy policies on "random subdomain/api/random string.html" or "random CloudFront domain/8-character random string.html". CineDream's volume came from paid acquisition (the app includes the AppsFlyer attribution SDK, and there is no organic content off-store). What drove it is a copy-and-paste playbook, not the product.
1. The data itself (googleplay library)
| Date | installs | Daily increase |
|---|---|---|
| 10-29 | 13,706 | – |
| 10-31 | 23,725 | +10,019 |
| 11-03 | 33,262 | +9,537 |
| 11-05 | 33,986 | +724 |
| 11-06 | 35,598 | +1,612 |
- 7 days +21,892, total 35,598 (public store tier 10K+). The two "+10k" jumps on 10/31 and 11/3 reflect Play's batch refresh of install counts. Other apps in the same matrix jumped on the same days (see below), so they shouldn't be read as two single-day spikes.
- Top Grossing: only in the Entertainment Top Grossing charts of Indonesia (id) and India (in), ranked between #160 and #190. Indonesia 10-29 170 at best #160 11-01, and later slid to #185/190 (chart_rank). Beyond the top 150 in Entertainment Top Grossing the bar is low, which means it has real paying revenue, but not much.
- No ratings and no reviews (ratings are empty in the library and review is 0). 35K installs with 0 ratings also points to one-time paid-acquisition users, not organic word of mouth.
- Version: latest on APKPure is 1.0.2 (2026-10-24,20.1MB), with the update note "minor bug fixes".
2. Competitors and "siblings": a reskin matrix
Searching the googleplay library by "privacy policy URL pattern" + "Mature 17+ + $2.99–$99.99" turns up apps from the same template (installs at 2026-11-06 or the most recent day):
| App | Developer account / legal name | Launched | Installs | 7-day gain | Best chart |
|---|---|---|---|---|---|
| HiFrame – Photo to Video AI | GE Technology / Ade Irma Setya Negara | 08-12 | 23,721 | 19,205 | id Entertainment Top Grossing #132 |
| Novora – Image & Video Maker | Unnica Telemetria… / PULIDO HERNANDEZ PATRICIA | 08-19 | 44,360 | 4,234 | mx Entertainment Top Grossing #101 |
| HoneyDrip – AI Video Magic | SupperSoul / Agung Setiawan | 08-20 | 90,486 | 39,321 | it Entertainment Top Grossing #80 |
| Sweetory – AI Video Generator | Frameverse / Shumaila Naz | 08-20 | 35,261 | – | id Entertainment Top Grossing #150 |
| Filmlet – AI Photo to Video | Lavish Food / Komal Jabeen | 09-02 | 53,342 | – | my Entertainment Top Free #140 |
| WowViz – AI Video Creator | WeDo Technologies / NINA MARLINA | 09-02 | 67,875 | 25,380 | id Entertainment Top Grossing #125 |
| CineDream – AI Video Magic | Zama Zama / ERIC LOPEZ | 09-03 | 35,598 | 21,892 | id Entertainment Top Grossing #160 |
| FilmForge: Photo to Video AI | Lavish Food / Komal Jabeen | 09-20 | 73,860 | 51,536 | in Entertainment Top Grossing #102 |
| LumiCraft–AI Video Experiences | Brady Carter / RANDALL JOHN BOWMAN | 09-21 | 32,321 | – | id Entertainment Top Grossing #119 |
| Aurion - AI Video Experiences | Perawat 24 / EDWIN ANTONIO HIDALGO MEJIA | 10-05 | 37,837 | 37,833 | my Entertainment Top Free #51 |
| VisionWave – Photo to Video AI | silambarasan / Asia Bibi | 10-14 | 556 | – | – |
| Woolora: AI Turns Vision Video | GLAD Technologies / THOMPSON ENOBONG | 10-16 | 26,623 | 24,497 | my Entertainment Top Free #68 |
Shared traits (all from the app table in the googleplay library):
- All Mature 17+, all Entertainment, all in-app purchase ranges $2.99–$99.99. Naming pattern: "X – AI Video Magic / Photo to Video AI / AI Video Experiences".
- Privacy policies follow two fixed patterns:
https://<random 10 letters>.<brand domain>/api/<random string>.html(CineDream is truclnerhpsu.play-cine.com/api/quhomadcdscq.html, HoneyDrip is ilxzmcbaof.honey-drip.com/…) andhttps://d<random>.cloudfront.net/<8 random characters>.html(newer builds since late September use this). - Developer legal names span Indonesian, Pakistani, Latin American, Nigerian and American names, with almost all emails personal Gmail. Developer names and legal names don't match (e.g. "Lavish Food" = Komal Jabeen, one account with two apps). This is typical of bought or rented developer accounts used to stack clone apps.
- Install counts jumped in sync on 10/31, 11/3 and 11/5 (app_day). New apps launched about every 1–2 weeks.
- The 12 apps in the matrix total about 520k installs. CineDream is only mid-tier.
- CineDream's description, beyond "photo to video," specifically says "virtual fashion / try on any outfit / Tap once, and your outfit is reborn," plus 17+ rating and "Users Interact." Guess: the templates include borderline outfit-swap/kiss-style effects (unverified; the store screenshots and description don't state it directly).
The real competitors (big players doing the same thing, googleplay library 2026-11-06):
- PixVerse (on 2024-09, ~80M installs, +7k in 520 days, #2 in Germany Photography Top Grossing)
- Nexus - AI Video Generator (DQL TechLab, 2026-312M, +7M in 1.85 days)
- Vidix (Supercool Labs, 2025-1.13B)
- Phant AI (2025-612M, #9 in Nigeria Entertainment Top Grossing)
- HubX "AI Video" (2024-316M), DancePop (2026-51.2M), Mitu AI Prank (4.8M)
- Web / big players: official apps for Kling, Hailuo, Pika, Vidu, Sora 2, Google Veo, plus the "photo come alive" effects built into CapTion/CapCut.
Compared with them, CineDream is two to three orders of magnitude smaller. The function is the same: upload a photo + pick a effect template → generate a video. No proprietary model (the privacy policy's boilerplate about "fine-tune stable diffusion" doesn't even match video generation).
3. Growth tactics
- Paid user acquisition (high confidence, specific channels unverified): the SDKs listed in the privacy policy are AppsFlyer (attribution) + Google Install Referrer + Google Ads Identifier/AppSet ID + Play Billing. That combination exists to measure ad ROI. There is no organic buzz off-store; 0k installs against 0 ratings from 35 reviews can only be bought traffic. Our Meta / TikTok / Google ad libraries don't show these package names (those libraries only track listed products, so absence doesn't mean they didn't advertise). From the charts, the spend looks concentrated in Indonesia and India; sister apps cover Italy, France, the UK, Mexico, Malaysia, Turkey, and others. Likely they test ROI per country by package.
- Account matrix / reskinned volume (confirmed): see the table above. The same template across 12+ packages under different accounts. When one package gets pulled or its ROI drops, they switch to the next, spreading the risk.
- Name piggybacking (weak): "CineDream" happens to be the name of a well-known color grading plugin (cinedream.io, widely used on TikTok and YouTube). Searching "CineDream" mostly returns the plugin. The Play search may catch a little leftover traffic, but it doesn't look intentional and isn't a major source.
- TikTok / X / Reddit / media: TikHub searches for "CineDream", "HoneyDrip AI", "FilmForge AI", and "WowViz" return nothing for these apps; x-radar has 60 posts in the last 0 days; on the web there are only auto-generated download pages like CNET, AppBrain, APKPure, and Softonic.
- Where the money comes from: one-time in-app purchases of $2.99–$99.99 (credit packs / weekly subscriptions). Price points are low for Indonesia and India, but Play CPIs there are also low. It has held a steady spot at #150–190 on Entertainment Top Grossing, which suggests the spend at least partly pays back.
4. Why it charted
- "Photo to video in one tap + borderline templates" has been shown for 2025–2026 years by PixVerse, Nexus, Vidix, and others to pay back paid acquisition. A template app only needs to plug into a third-party video model API (timing: video generation APIs are getting cheaper).
- The team turns the whole setup into a repeatable template: batch-swap names and icons, publish under different developer accounts, and give each app a small test budget. Whatever runs gets more budget (distribution playbook).
- Paid acquisition + AppsFlyer attribution, targeting low-CPI markets like Indonesia and India, and recovering spend from a small share of payers through high-priced in-app purchases (the $99.99 tier).
- The 17+ rating and the outfit-swap hint likely boost click-through (guess, unverified).
5. Differentiation
No product-level differentiation. Next to big players like PixVerse, it has no proprietary model, no brand, and no community. Next to its sister apps in the matrix, it's the same app with a different name and icon. Its only "edge" is the playbook: spreading ad risk and takedown risk across many small pieces with an account matrix. It works because of that playbook, not because of the product.
6. Visibility and copycats
- Visibility: almost nobody knows about it. Revenue isn't public; there's no press, X, or TikTok discussion, only auto-generated APK download pages.
- Copycats: it is itself one member of a copycat matrix. Beyond this template, the library also has Framegen (SecureVault, 09-22) and Afterglow, among other 17+ similar photo-to-video small apps.
7. What others can learn
- The "many accounts × same template × per-country ROI testing" playbook lets a simple wrapper app make money from paid acquisition. One dead package doesn't hurt the whole operation.
- Test first in low-threshold, low-CPI entertainment markets like Indonesia and India.
- The risks are also direct: Google Play can ban linked accounts, 17+ borderline content, and a sloppy privacy policy (boilerplate SD avatar text). Account costs are a continuous drain.
8. Uncertain
- Which specific channels are used (Google UAC, Meta, or networks like Mintegral/Unity) is unverified; the ad libraries don't cover these packages.
- Whether the app actually contains borderline or outfit-swap effects is unverified (not installed to check).
- Who runs the matrix (guess: one team or a user-acquisition studio) is unverified; the registered developer names clearly look bought.
- Actual revenue is unknown. We can only see its #160–190 rank on Indonesia / India Entertainment Top Grossing.
Sources
- googleplay library (agent-im, 2026-10-29 to 2026-11-06): app, app_profile, app_day, chart_rank
- Play Store page: play.google.com/store/apps/details?id=com.generator.cinedrea…
- Privacy policy (includes SDK list): truclnerhpsu.play-cine.com/api/quhomadcdscq.html (last updated 2026-08-21)
- APKPure version page: apkpure.com/cinedream-%E2%80%93-ai-video-magic/com.generator…
- Same-name plugin: www.cinedream.io/
- TikHub TikTok search, x-radar search (2026-11-06, nothing relevant); metaads / tiktokads / googleads libraries (package name not found)