# A 16-second ad ran for 45 days, and installs stopped the day it was paused: Vidfig was a failed test · Mobile app

> Source: https://stringers.io/en/apps/play_com.vidfig.app.html (Stringers / 什么值得做, generated 2026-11-08)
> Stringers answers one question every day: what small product can one or two people build in a few months and actually make money from. Numbers come from 15 databases we collect daily, each with source and date.
> For agents: these verdicts and the underlying databases (verified revenue, App Store / Google Play charts, ad libraries, GitHub, Product Hunt…) are queryable live over MCP — tools today / hot / products / directions / direction / data. Setup: https://github.com/komako-workshop/stringers (the user logs in at stringers.io → "配置 agent" to get a personal URL). Site index: https://stringers.io/llms.txt

> It hasn't really taken off. A new Hong Kong company released three shells on the same backend. Vidfig handled US testing, and a 16-second Google video ad ran for 45 days and brought in most of the installs. The category Top Grossing chart has a low bar, so a few thousand installs was enough to get it onto the chart.

[Mobile App](https://stringers.io/en/apps/index.md) · Android · written 11-6

Vidfig: AI Video Maker · Tyrone Hub

# A 16-second ad ran for 45 days, and installs stopped the day it was paused: Vidfig was a failed test

**What it is** An Android shell that calls third-party video models: upload a photo or reference video, generate a few seconds of motion, pay per credit pack ($3.59–77.99).

**Why it took off** It didn't really take off: a new Hong Kong company used one backend to ship three shells, Vidfig was the one testing the US, and a single 16-second Google video ad that ran for 45 days brought most of the installs. The bar for the category's top-grossing chart is low, so a few thousand installs can put it on the chart.

**D** **How you could follow it** Don't build another image-to-video shell for the US. Teams already running this type of shell can borrow its testing method: one shell per country, keep one of two or three creatives running for over a month, and decide whether to add budget or stop based on payback.

Grade D: no product differentiation, the model belongs to someone else, and the US charts already have Kling, Dreamina and dozens of bigger shells. The team behind it has a backend and knows paid acquisition, but after 45 days of US spend they pulled out and moved to other countries. Almost nobody knows about it yet

Data (updated daily) · first launch 2026-08-22 · 11-07 installs 5,994

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.

Android · Ratings count (cumulative)

Data 2026-10-29 → 2026-11-07 (10 days, 10 observations)

Vidfig is an AI image-to-video app on Google Play. It has been listed for two and a half months, has 6 thousand installs, and sits at #78 on the US Art & Design top-grossing chart. The numbers look like it is picking up momentum, but nearly all of its new installs were bought by a single Google video ad: the ad stopped running in Nov 2, after which daily new installs fell from one or two hundred to 12. This piece explains who made it, where the traffic came from, how it charges, and why you shouldn't enter this track again.

### One ad from a gym

In Sep 18 of 2026, a 16-second vertical video began running in Google ad placements in the US: in a gym, a short-haired girl in a grey sports tank smiles at the camera, with an AI-generated person and background. On YouTube the video was titled "Create Any Video You Like 05", and tapping it led to Vidfig's store page. The ad ran continuously for 45 days, and the last impression was in Nov 2 (Google Ads Transparency Center, 11-06).

In early September after launch, it tested two ads at once: "Create Any Video You Like 02" (from 9-01) and "Turn photo into any video" (from 9-10). When the main creative went live on Sep 18, those two were stopped the same day. All three ran in the US, and the one that survived was the gym girl.

### What it is

Upload a photo or a reference video and generate a few seconds of motion: animate a person, regenerate according to the motion of a reference clip, adjust styles, export in portrait or landscape. The store listing doesn't say which model it uses, but a sister app under the same developer account explicitly says "Seedance," so it is likely also calling a third-party model like ByteDance's Seedance (inferred).

Free to download, monetized through in-app credit purchases priced at $3.59 to $77.99 each. The free credits for new users are just enough to make one video. A US user wrote in Sep 9: "I made 1 videos, and it really is like the ad says: upload a picture, type one sentence, and it starts generating. The starting credits are enough for one 4-second video... if I can get more credits I'll try more complex prompts." Another said: "The templates are nice, the results are hit or miss, but I see a lot of potential. I'd suggest spending a few dollars to try it." Store rating based on 4.4,82 reviews (Google Play, 11-06).

### How the traffic came in, and how it went away

| Date | Total installs | Notes |
|---|---|---|
| 09-25 | 1,783 | Main ad ran for one week (third-party estimate) |
| 10-03 to 10-09 | About +220 per day | Fastest stretch (third-party estimate) |
| 10-31 | 5,517 | Google Play |
| 11-03 | 5,894 | +377 versus last time; the day after the ad stopped, the number still reflected the previous days' volume (store figures lag by a day or two) |
| 11-05 | 5,982 | +88 |
| 11-06 | 5,994 | +12 |

During the 45 days the ad was running, it averaged over a hundred installs a day; once the ad stopped, that fell to 88, then to 12. At the same time it dropped off the US Art & Design free chart (where it had previously sat between #114 and #164). Meta and TikTok ad libraries show no sign of it, a TikTok search for "Vidfig" turns up no related videos, X has 0 posts over 90 days, and the web only has auto-indexed pages from download sites. No organic traffic: paid acquisition is everything.

Don't take that #78 top-grossing ranking at face value either. Of the top 80 on the US Art & Design top-grossing chart, at least 25 are AI image-to-video apps, but the bar for this category is very low: a few thousand installs and a handful of credit-pack purchases can put you at #60 to #80 (Google Play category chart, 11-06).

### Who made it: one company, three shells

The developer is called Tyrone Hub. The advertiser is Fortune Horse Technology Limited (駿澤科技) in Hong Kong, registered in May 2026 (15). Its privacy policy and support email point to it. But Play's "About the developer" lists a different company, Tyrone Labs Limited, registered in Sep 29, which is even later than Vidfig's release. Keeping the operating company and the developer account entity separate is a common way for global-facing tool teams to isolate ban risk. One more detail: the ad videos all sit under a YouTube channel called Litestartup, a platform for startups offering email, site building and a unified AI model API. It could be the same team's other project, or just borrowed (inference, not verified).

One account, three nearly identical apps in just over two months, each running ads in different countries:

| App | Launched | Installs (11-06) | Main market | Best rank |
|---|---|---|---|---|
| Pixvi | 06-21 | 20,540 | Japan | Japan Art & Design Top Grossing #11 |
| Vidfig | 08-22 | 5,994 | US | US Art & Design Top Grossing #63 |
| VidImg | 08-30 | 11,190 | Commonwealth countries, South Korea | Japan Personalization Top Grossing #3 |

Vidfig is the smallest of the three and the only one aimed at the US. Its ads stopped in Nov 2, and in Nov 3 VidImg's ads started running in South Korea. The budget moved, so Vidfig is most likely the one that didn't pencil out (inference).

### Who it's competing with

The US is the most crowded market for this category. On the US Art & Design Top Grossing chart in Nov 6:

| Competitors | Installs | How it differs from Vidfig |
|---|---|---|
| Kling AI | 34.8M, Top Grossing #2 | Model makers' own apps |
| Runway | 2.56M, #3 | Proprietary models, aimed at professional creators |
| AIReel | 5.03M, #5 | Same kind of wrapper app, hundreds of times the scale |
| Dreamina (ByteDance) | 6.02M, #39 | Official sale of Seedance directly |
| SeeGen AI: Seedance 2.0 & 2.5 | 110k, #35 | Model name in the title, which captures store search |
| Pollo.ai, Vidful.ai, etc. | Web | One site bundling Kling, Veo, Seedance |

The models belong to someone else, and the model makers' own apps are on the chart, along with dozens of clones of the same thing. Vidfig has nothing product-wise that others don't, and it ranks in the last tier on scale.

### If it were you

**D. Don't follow.** If you build another AI image-to-video wrapper for the US, you're up against Kling and Dreamina's own apps and dozens of wrappers whose paid acquisition is hundreds of times yours. Vidfig's team had a ready backend and knew how to run ads, and after 45 days of US spend they still pulled out.

**Who it suits** : almost nobody. If you have to pick someone, it's a going-global team that already has a video generation backend, several wrapper apps, and wants to test cheaply which country pays back, and can borrow its approach to check their own workflow.

**Who it doesn't suit** : indie devs who want to make money from the product; anyone without paid acquisition experience who expects organic store installs (two days without spend and its new installs drop to zero); anyone who wants to focus on the US market.

**How to do it (if you're already building this kind of wrapper)** : what's worth copying is the testing cadence, not the product. One backend, several brands; each wrapper goes after one main-market country. Run two or three 16-second vertical creatives at once, keep one after a week or two, run the winner continuously for over a month, and decide whether to add budget or stop based on returns. Give new users only one 4-second video's worth of credits so they hit the paywall right after seeing the result. When reading install numbers, remember the store figures lag by a day or two, and the effect of stopping spend takes two or three days to show.

**How it acquires users** : it proves one thing. Apps like this have no organic traffic, and Google App Campaigns are the only real source. Creatives use AI-generated "good-looking people moving," directly demonstrating what the product can do.

**When to move before** : there's no window. The model makers' official apps keep improving and giving away more free quota, so this category will only get more crowded.

**What would show the read is wrong** : ① Vidfig reappears with new US creatives in November and daily installs recover to hundreds, meaning 11-02 was just swapping creatives, not giving up, and payback may be better than it looks; ② it keeps growing installs steadily without ads, meaning there's organic traffic we can't see from outside; ③ on the US Art & Design Top Grossing chart, a small app with a few thousand installs holds a top 50 spot for a long time, meaning paying users here are thicker than they look.

---

Data and sources: Google Play store page (US store, read 11-06); installs and category rankings (10-29 to 11-06); installs from 9-25 to 10-09 are third-party estimates; Google Ads Transparency Center (advertiser Fortune Horse Technology Limited, first and last ad impression dates, as of 11-06); YouTube video pages (ad creative titles, durations); Hong Kong Companies Registry new-incorporation lists, ltddir, hkregister; Meta and TikTok ad libraries and TikTok and X searches (11-06). Where it says "inference" or "estimate," it is an inference or estimate.

[Android store page ↗](https://play.google.com/store/apps/details?id=com.vidfig.app)

**Research notes: competitor table, acquisition evidence, full research report**

Competitors

| Name | Where | Launched | Scale | How it differs from this one |
|---|---|---|---|---|
| Kling AI | Android | 2024 | 34.8M installs, US Art & Design Top Grossing #2 | A model vendor's own app with a big brand; Vidfig is a small wrapper calling third-party models |
| Runway | Android | 2024 | 2.56M installs, #3 on Top Grossing | Professional creator tools, proprietary models |
| Dreamina (ByteDance) / Vidu (Shengshu) | Big-company features | 2024–2025 | 6.02M / 2.33M installs | Big-company apps with Seedance / Vidu models built in, with better price and capability |
| AIReel(FunCreate) | Android | Not verified | 5.03M installs, #5 on Top Grossing | Same kind of wrapper app, but templates and paid acquisition at hundreds of times the scale |
| MePic / Lumy / Couple AI(Ocean Float Mobile) | Android | Not verified | 8.59M / 790k / 150k installs | Same setup, one company with several apps, but much bigger scale |
| SeeGen AI: Seedance 2.0 & 2.5 | Android | Not verified | 110k installs, Top Grossing #35 | Title rides the Seedance model name directly to capture store search traffic |
| Revive(REFACEAI) | Android | Not verified | Not verified | Veteran photo-animation app; Play lists it as a similar app to Vidfig |
| Vidful.ai / Pollo.ai | Web | 2024–2025 | Not verified | Web aggregator combining several models, acquiring users through SEO |
| Pixvi / VidImg (same company) | Android | 2026-06-21 / 2026-08-30 | 20.5k / 11.2k installs, Japan category Top Grossing #11 / #3 | One backend, different names and different main-market countries |

Where users come from

- **Verified** **Google Ads video install ads (US)** : per the googleads library, advertiser Fortune Horse Technology Limited ran 3 Vidfig creatives, all targeting the US. The main creative ran from 9-18 to 11-02 continuously for 45 days. After spend stopped, new installs dropped from hundreds a day to 12.

- Inference **Organic store traffic / category charts** : a small amount of exposure from ranking #114–164 in the US Art & Design free chart and #63–78 in the top grossing chart; after spend stopped it fell off the free chart.

- Inference **Meta / TikTok / creators / X** : nothing found in the metaads and tiktokads libraries, nothing in TikHub search, x-radar 0 posts.

## Vidfig: AI Video Maker (com.vidfig.app) research report

> Channel: Google Play | Developer account: Tyrone Hub | Listed: 2026-08-22 | Research date: 2026-11-06

### Bottom line

Vidfig didn't "take off" on its product. It's one of three AI image-to-video wrapper apps (Pixvi / Vidfig / VidImg) built on the same backend by a newly registered small Hong Kong company, Fortune Horse Technology Limited (駿澤科技, founded 2026-05-15). It **got its installs from Google Ads video acquisition** . It ran only in the US, with one main creative, and installs stopped when the ads stopped. The Art & Design Top Grossing chart in the US ranked it #63–78, a low bar since AI video wrappers crowd this category. That ranking suggests some paying users, but a very small scale, and not a unique opportunity.

### 1. Basic data (googleplay library, 2026-10-29 to 11-06)

| Metric | Value |
|---|---|
| installs | 5,123(10-29) → 5,994(11-06), +871 in 7 days |
| Score | 74 → 81, store rating 4.4 |
| Charted | US ART_AND_DESIGN Top Grossing #63–78 (best ever #63,10-30); same category in the Netherlands Top Grossing #138–173; US same category Top Free #114–164, **dropped off the free chart after 11-03** |
| In-app purchases | $3.59–$77.99 each (credit packs or subscriptions); reviews say the free credits for new users are only enough to generate one 4-second video |
| Historical installs (app_day_ext, backfilled by AppstoreSpy) | 9-25 cumulative 1,783 → 10-09 3,593 (10-03~10-09 about 220 a day) → 10-31 5,517 → 11-02 5,894 → 11-05, now only 12 a day |
| Recent reviews | None in the library. The few on the Play page are English praise from September ("worked as advertised" "good templates, outcomes are iffy"). |

### 2. Who's behind it: one company, three apps

- The developer on Play is called "Tyrone Hub." The support email and privacy policy both point to **greatfortunehorse.com** , and the privacy policy is signed by Fortune Horse Technology Limited. The Hong Kong Companies Registry's new-company list shows this company was incorporated in 2026-05-15, company number 80411213 (cr.gov.hk RNC063, from the 2026-05-11~17 batch; databasesets.com). Its website 2026-05-28 went live, describing itself as "AI-powered creative products… subscriptions, credits, cloud processing, advertising."

- The "About the developer" section on the Play page lists **Tyrone Labs Limited** (泰宸龍宇科技, corp@tyronehub.com, address in Tsim Sha Tsui). That company was **incorporated 2026-09-29** (ltddir.com, hkregister.com), even later than Vidfig's release. It appears to have been registered specifically to hold the developer account's entity information, separate from the operator Fortune Horse. This is a common pattern for going-global tool apps: multiple entities and multiple accounts, to isolate risk.

- Three apps under one account (googleplay library): | App | Listed | Installs | 7-day growth | Best rank | |---|---|---|---|---| | Pixvi: AI Video & Image | 2026-06-21 | 20,540 | +1,104 | Japan ART_AND_DESIGN Top Grossing #11 | | VidImg: AI Photo & Video (description says it "leverages the advanced Seedance AI model") | 2026-08-30 | 11,190 | +1,367 | Japan PERSONALIZATION Top Grossing #3 | | Vidfig: AI Video Maker | 2026-08-22 | 5,994 | +871 | US ART_AND_DESIGN Top Grossing #78 | All three do "upload a photo or reference video → generate a video," with nearly identical selling points. My best guess is they call the same third-party video model (a Seedance-type model), changing only the name, category and main-market country. This is a guess, not verified.

### 3. Growth tactic: Google Ads video acquisition (confirmed)

googleads library, advertiser **Fortune Horse Technology Limited** (AR00290949683155566593, entity verified):

- 10 video creatives in total, all install ads linking to the app. The earliest started running on 2026-07-02, before Vidfig launched, so it was probably Pixvi.

- **Vidfig had 3 creatives, all targeting the US** : 9-01~9-18, 9-10~9-18, plus a main creative **9-18~11-02** (run continuously for 45 days, US + GE).

- VidImg's creatives run in GB, CA, DE, IL, NZ, AU (9-06~10-25), and it started running ads in Korea on 11-03.

- **Ads stopped, installs stopped too** : Vidfig's main creative was last shown on 11-02. Store installs went from 11-02 to 11-03 and still rose by 377 (the store number lags), then +88 and +12 per day afterward. The same day it also dropped off the US Top Free chart in Art & Design. This mostly suggests nearly all its new installs came from paid acquisition.

- Neither the Meta Ad Library nor the TikTok Ad Library turned anything up (Meta doesn't index this advertiser; on TikTok, VidImg is listed as no_ads). Searching "Vidfig" on TikTok returns nothing relevant, and on X, both Vidfig and Pixvi have 0 posts. Online there are only auto-indexed download-site pages from Softonic, CNET Download, and AppBrain, no press coverage and no influencer promotion.

- Where the money comes from: in-app credits, up to $77.99. New users get credits for only one 4-second video, so they have to pay once they've used it up. The US Art & Design Top Grossing chart has a low bar: a few thousand installs can land you at #60–80. Going by category rank, ROI was probably break-even to slightly profitable (not verified). Stopping the spend could mean they didn't work out the numbers, or that budget moved to VidImg (Korea) and Pixvi (Japan).

### 4. Competitors (the category is extremely crowded)

2026-11-06 In the US ART_AND_DESIGN Top Grossing top 80 (googleplay library), there are at least 25 "AI image-to-video" Apps:

- Big-company or model-maker apps: Kling AI (34.8M installs, Top Grossing #2), Runway (2.56M, #3), Dreamina / ByteDance (6.02M, #39), Vidu / Shengshu (2.33M, #46), Adobe Firefly (3.62M, #79).

- Going-global shell portfolios: AIReel (FunCreate, 5.03M, #5), VifunAI (550K, #8), MePic / Lumy / Couple AI (three apps from one company, Ocean Float Mobile), Drma (4.86M), DeeVid (740K), SeaArt (7.85M), SeeGen "Seedance 2.0 & 2.5" (110K, #35), plus Anijam, Movify, Pixaflow, Storyla, Vidoza, Reelify and others. All launched this year and sit at tens of thousands to hundreds of thousands of installs.

- Similar apps Play recommends alongside Vidfig: Revive (REFACEAI), Vidu, Drma, Monet (APPNATION).

- On the web: aggregator sites like Vidful.ai, Pollo.ai and vidflux.ai bundle Kling, Veo and Seedance and resell them.

**Vidfig is the smallest tier in this group** (6K installs), and its features match everyone else's.

### 5. Differentiation

Almost none. "Reference-video-driven motion" is a capability of models like Seedance or Kling, and every shell in this category has it. Its real "edge" is the playbook: **one backend feeding multiple shells, each targeting one country** (Vidfig targets the US, VidImg targets the Commonwealth and Korea, Pixvi focuses on Japan), testing each shell's ROI with Google Ads and adding money to whichever works. So it getting picked up is due to the ad buying, not any product difference.

### 6. Visibility and copycats

- Visibility: **few people know about it** . No public revenue, 0 mentions on X, no press coverage.

- Copying: hard to say who copied whom. It is itself chasing a category that was already packed (AI image-to-video shells). The same company's other two shells, Pixvi and VidImg, are doing better than it is.

### 7. What to borrow

- One model API backend, multiple brand shells, listed separately by country or category. Top Grossing charts in small countries and small categories have low bars, so a few thousand installs can get you charted, which makes it easy to test paid conversion.

- Google App Campaign video creatives only, one main country per app, one lead creative running for about 45 days, then decide whether to stop or scale based on ROI.

- New users get credits for "one 4-second" video only, so the effect hits them and they immediately hit the paywall, with credit packs starting at $3.59.

- The developer account entity and the operating company are kept separate (Tyrone Labs and Fortune Horse), to isolate ban risk.

### 8. Uncertain

- Actual revenue and ROI are unverified; it's unclear how much revenue corresponds to US Top Grossing #78.

- Whether the underlying model is Seedance: VidImg's description says so, Vidfig's doesn't. This is a guess.

- Channels beyond Google (Meta, etc.) simply aren't in our database. That doesn't mean they weren't used.

- Whether 11-02 stopping spend means abandoning this shell or rotating creatives will depend on whether new creatives appear in the next few days.

### Sources

- googleplay library: app, app_profile, app_day, app_day_ext, chart_rank (2026-10-29~11-06)

- googleads library: advertiser, creative, adv_app, advertiser_snap (advertiser Fortune Horse Technology Limited)

- metaads / tiktokads libraries: nothing found

- Play page: [play.google.com/store/apps/details?id=com.vidfig.app](https://play.google.com/store/apps/details?id=com.vidfig.app)

- Official site: [www.greatfortunehorse.com/index.html](https://www.greatfortunehorse.com/index.html)

- Hong Kong company registration: [www.cr.gov.hk/docs/wrpt/RNC063_2026.05.11-2026.05.17.pdf](https://www.cr.gov.hk/docs/wrpt/RNC063_2026.05.11-2026.05.17.pdf); [www.ltddir.com/companies/tyrone-labs-limited/](https://www.ltddir.com/companies/tyrone-labs-limited/)

- Download site: [download.cnet.com/vidfig-ai-video-maker/3000-android-vidfig-…](https://download.cnet.com/vidfig-ai-video-maker/3000-android-vidfig-ai-video-maker.html)

- X-radar search for "Vidfig" / "Pixvi" (90 days) returned 0 posts; TikHub search for "Vidfig" returned no relevant videos.

This article was written by AI from public data and the data we collect daily. Numbers are sourced, and inferences are labeled.