An app that only "copies": 590k installs in ten weeks, a Bangladeshi AI video prompt library

A small Bangladeshi developer built an app with no AI in it at all, just over a thousand AI video prompts. You watch the sample, tap to copy, then generate in Google Flow. Ten weeks after launch it has 590k installs, nearly all growth coming from South Asian short-video creators making "viral effect" tutorials. Below: how it got started, where the money comes from, who's competing, and whether you could do it again in another country or with another generation tool.

A 65k-view Reel

On the afternoon of Sep 22, 2026, an Indian Facebook account named Anjani Ratan posted a vertical video titled just "Try this viral edit ✨👀," with a misspelled hashtag #cartreandingvideo (car trending video). Going by the search summaries from that period, these videos all follow the same formula: start with an AI-generated "driving" viral clip, then demonstrate copying a prompt from Easy Prompt and pasting it into Google Flow. By November, the video had 65k views and 1 thousand likes. The account's other videos mostly had only a few dozen views.

The first comment was in Hindi: "I can't make this work at all, tell me why the video won't download" (Sadik Ali). The people watching want that effect, and the missing step, "where do I get the prompt," has one answer: an app.

Three days after Sep 25, Easy Prompt gained about 55k more installs in a single day (AppstoreSpy estimate). I can't match this to a specific video, but its growth curve looks like this: flat for a few days, then a sharp one-day spike, which fits the pattern of a tutorial video going viral and lifting it for a while.

What it is: a prompt catalog for AI video

Easy Prompt - Ai Video Prompt, developer Sajjad IT. The store page lists Md Sajjadur Rahman in Chandpur, Chandpur district, Bangladesh. Published on Google Play in 2026 Aug 29, category Art & Design. No iOS version.

The app itself is thin: it lists AI video samples by category, lets you view the prompt, and copies it with one tap. The store page sells three things: "Preview Before You Copy", "One-Tap Copy", and "1000+ free prompts, daily prompts uploaded". It generates nothing. The bottom of the store page even says it is "not affiliated with any AI video generation platform". Users copy a prompt and generate the video themselves in Google Flow (Veo) or Gemini.

This developer isn't new to this. In 2025 November he published GoPrompt - Ai Image Prompt, the same model but for image prompts (at the time, Nano Banana photo editing in Gemini was popular in South Asia), and by early November it had 2.27M installs. Easy Prompt takes the same approach and swaps images for video.

Where the traffic comes from

Its home market is Bangladesh. In Nov 6 it ranked #3 on the Bangladesh Art & Design Top Free chart and #193 on the overall Bangladesh Top Free chart (AppWinners), and charted in 90 countries in total. The other strong cluster is the Gulf: #21 in Qatar, #32 in Saudi Arabia, #36 in the UAE. There are many Bangladeshi and South Asian migrant workers there, so the same group is likely installing it (inference).

The traffic mainly comes from South Asian Facebook Reels and short-video tutorial accounts, not from the developer:

• Indian account Anjani Ratan, a viral "car" tutorial from 9-22 with 65k views.
• Bangladeshi account Tech Jani posted "হাজার হাজার AI Video Prompt এক অ্যাপে!" (thousands of AI video prompts in one app) 9-29, which got 1.2K views. The account usually posts phone tips like "turn this Chrome setting off now".
• You can also find several similar tutorials such as "Easy prompt video ai new trending car video" and "Trending AI Walk Couple Video Editing Prompt". They all follow the same loop: viral effect → get the prompt in the app → paste into Google Flow.

The developer's own accounts don't explain the 590k installs: YouTube @sajjaditdotcom has 75.6k subscribers, but its most recent long video was about GoPrompt 11 months ago; TikTok has 1,244 followers, and his September Google Flow tutorials got a few hundred to over a thousand views; the blog sajjadit.com publishes SEO posts like "[X] AI prompt, copy and use". These look more like raw material and templates for tutorial creators than a traffic engine.

The other three sources aren't confirmed: whether GoPrompt's 2.27M older users cross-promote it (inferred, not proven); whether it runs Meta ads (one fake-name page 10-24 posted a Bengali-language video promoting "Easy Prompt AI" as a real person, but it talks about photos, so it's unclear if that's them); and putting "Ai Video Prompt" straight into the title, since people searching that phrase on Play can easily land on it.

Why now

In 2026, Google Flow / Veo caught on in short-video circles in Bangladesh and India, and "making viral videos with Google Flow" became its own tutorial category. The developer himself posted "80's Trending Video Tutorial #GoogleFlow #VideoPrompt" on TikTok in Sep 14. The generation tools are free or cheap; the hard part is writing a long English prompt. For users with weak English on low-end Android phones, "copy a ready-made one" is the whole need.

The Nano Banana photo-editing wave a year earlier had already trained the same users to "go to an app for prompts". GoPrompt and many other image-prompt apps took off then. Video is the next step for the same people, and Easy Prompt is one of the few that targeted that step specifically.

Where the money comes from

Ads only. The store lists "Contains ads" and there are no in-app purchases. A three-star review from Sep 27 (Nayeem Bhai, 6 people found it helpful) says: "Only three stars because I have to watch extra ads... if you change that, more people will download it." Inference: you have to watch rewarded ads to copy or unlock prompts.

Revenue is not public. Rough estimate: Bangladeshi ad rates are very low, and most of the 590k installers probably copy one or two prompts and leave. Ratings are only about 0.1% of installs (GoPrompt is 0.7%), which suggests one-time, light users. At a few cents to a cent or two of ad revenue per install, the ten-week total is probably a few thousand to ten-odd thousand dollars (estimate, low confidence). For an individual developer in Chandpur that is decent money; for a team elsewhere it is a thin, volume-driven business.

Who's competing

The image-prompt category on Play was already crowded, nearly all from small developers in India, Pakistan, and Bangladesh:

As far as I can tell, Easy Prompt is currently the only app with "video prompt" in its title that has reached meaningful scale. But another app named "Easy Prompt" (com.lsiappstudio.easyprompt) launched at about the same time. It ran a promo video on Facebook in Aug 25, focused on images, and its scale is unverified. English-language video prompt libraries (youmind, promptessor, and similar) target creators in Europe and the US and don't reach South Asian phone users. The category hardly exists on iOS, and comparable US-store apps have only a few dozen ratings.

It has no moat. The content is public prompts that anyone can copy, and the distribution lives in other people's Reels, not in the app.

If it were you

Grade B. You could do it, but this is a business that follows the hype of generation tools, relies on local tutorial accounts for distribution, and earns thin ad revenue. It is not a product opportunity. What is worth learning is the playbook, not the app.

Who it suits: a small team of 1–2 people who can make short videos in an emerging-market language outside South Asia (Indonesian, Vietnamese, Urdu, Arabic, Tagalog, Brazilian Portuguese), who know or can pay local "AI tips" creators, who can integrate AdMob rewarded ads, and who live somewhere with low costs, where a few thousand dollars a month counts as success.

Who it doesn't suit: anyone who wants subscriptions or high ticket prices (these users don't pay); anyone who only writes code and doesn't work with short video (all the traffic is in Reels); anyone who wants to head-on copy a video prompt app in Bangladesh or India (the developer is already established and same-name copies have appeared).

How to do it: change the place or the tool, not the playbook. First, change markets: pick a country where Google Flow / Veo is just starting to trend in local short video but no local-language prompt app exists yet, and build a local-language video prompt catalog with a sample for each entry, one-tap copy, and rewarded-ad unlocks. Second, change the upstream tool: whichever generator is taking off in a given market (Grok Imagine, Seedance, Kling), build a "prompt companion" for it. Update prompts daily, featuring only what's viral locally right now, rather than a generic "cinematic" encyclopedia.

How to get users: run 3–5 local-language "viral effect tutorial" Reels / TikTok accounts, each covering one effect and ending with "the prompt is in the app". At the same time, give prompts free to local AI-tips creators so their tutorials point to your app. Watch the install curve: if it jumps, find out which video drove it, then make ten more like it. Put the term locals actually search for in the title.

Deadline: ship the first market by end of Dec 2026. This category rides the generation tools' hype. Easy Prompt went from launch to its single-day peak in five weeks and daily adds had fallen to a few thousand by mid-October. Once the "Google Flow viral video" buzz in a market fades, the window closes.

How I could be wrong: ① Easy Prompt's weekly new installs fell below 10k in December, and GoPrompt is also declining, which suggests the whole prompt-app wave is receding, not just one market saturating. ② Google Flow or Gemini may ship built-in, region-specific "trending templates" you can apply with one click, which would eliminate this middle layer. ③ Google Play could remove prompt-only apps for "thin content / limited functionality" (inference: this risk exists, not yet seen). ④ The market you pick could see two or three local-language video prompt apps reach the top 20 of its category within three months. Then you're late, so move to the next market.

Data and sources

• Installs, ratings, and charts: Google Play store page and chart data from 30 countries, as of 11-06. Daily curves since launch are AppstoreSpy estimates. Bangladesh and Gulf-country rankings come from AppWinners (11-06).
• Developer info, store description, and reviews: Google Play store page (Bangladesh storefront, 11-06). The Bangladesh storefront shows 4.5 stars and 607 reviews; the global figures are 4.2 stars and 602 reviews.
• Tutorial videos: public Facebook videos from Anjani Ratan (9-22), Tech Jani (9-29), Easy Prompt Apps (8-25), and Winters Smallkmi (10-24). View counts as seen: 11-06.
• Developer's own accounts: YouTube @sajjaditdotcom, TikTok @sajjaditdotcom, sajjadit.com.
• Comparable apps' scale: Google Play, as of 11-06.
• Revenue figures are estimates. The developer has not published any revenue numbers.