Take a photo and AI tells you how much it weighs: a Turkish indie developer brought the "phone scale" flood from iOS to Android

Scale for Grams: AI Estimate is an Android tool built by an indie developer in Istanbul. You point your camera at an object and a large model estimates how many grams it weighs. Its features are indistinguishable from the hundreds of same-named iOS apps. It gained traction through three things: a title and search terms copied word for word, launching on Android, and growing in countries with a low bar for charting: Indonesia, Thailand, and Mexico. It adds about 12k installs a week, which is not big. What's worth learning is the playbook of switching platforms and switching countries. The product itself isn't worth copying again.

A 39.1-gram ring was estimated at 6 grams

On Nov 5, 2026, a user left this review on Google Play: "Said a 39.1gram keeper ring was 6gram beaded ring 🤣🤣🤣". A 39.1-gram ring, and the app said it was a 6-gram beaded ring, off by more than six times.

Coincidentally, the app's website homepage uses a gold ring as its demo: the camera frames the ring, and the card reads "about 9.07 grams", with a history below it showing a gold chain, coffee beans, and car keys. What users point their phones at is what the developer wants them to point at. Whether the estimate is accurate depends entirely on the model's luck.

The review didn't stop it. Over the same week (11-05, Google Play installs), its cumulative installs rose from 10-29 to 22,427 and then to 34,141,7, up about 11,714 days over, a gain of roughly 52%; rating count went from 301 to 433.

What it is

A photo-based weight estimator. Take a picture or choose one from your gallery, select "estimate weight", and after a few seconds you get a gram value, a confidence level, and a few lines of explanation, with metric and imperial toggles. It also bundles 5 other photo tools: length estimation, counting, calories, plant identification, and photo translation.

The store description has a long disclaimer: "This app does not turn your phone into a physical or certified scale", and it says not to use it for medical, laboratory, mailed-goods pricing, or trade purposes. The store subtitle also says "informational, not a real scale".

Money comes from two places: the store page lists "Contains ads" and "In-app purchases", with in-app purchases at $2.99–$39.99 each; the listing says the paid version "allows more analyses and a stronger AI". Sales volume isn't public. It has never charted in any country's Top Grossing, so revenue is very likely small (inferred).

The developer is HAMZA KAVAK, publishing as Glipo Devs, based in Istanbul. He also has photo-based apps for counting, plant identification, PDF scanning, and calories, and this app essentially combines them. On iOS he released a similar "Weighing Scales: Gram & Weight" in 2026 March, which didn't take off. He tried again on Android in Jul 24, and this is that version. More than three months after launch, the release notes still say "Welcome to the first release!", with little iteration.

How it took off

It started with a very slow cold start. According to a third-party estimate (AppstoreSpy), it had only 57 cumulative installs by Aug 12, dozens per day in September, and 1,489 in total by Sep 16.

From late September it climbed in steps. Oct 4 single day +748, Oct 7 single day +2,789, then Oct 13 through 27 at 600 to 1,500 per day. Several days in between were close to zero: 10-05 +2,10-16 +7,10-28 +18. Organic store-search traffic is usually smooth, and this stepped shape looks more like ad budgets switched on and off (inferred; no direct evidence of its ad spend was found).

It charted in only one category: Health & Fitness Free, and only in three countries. Between 10-29 and 11-06, its best rank in Indonesia was #86 (climbing up from #100 little by little), in Thailand it sat at #153–#175, and in Mexico at #185–#199. From 11-01 to 11-03 it briefly appeared at the bottom of the Argentina and Spain charts. It never charted in the US, on the overall chart, or in Top Grossing (Google Play charts).

Taken together, these points show its approach:

• The title is the search term. "Scale for Grams" is the most common English search for this need, and it uses it verbatim as its name. The website title "Online & Phone Scale App" draws on a different term, "phone scale".
• It picked Health & Fitness, not Tools. In the Health & Fitness category charts of Indonesia, Thailand, and Mexico, roughly a thousand installs a day is enough to reach the top 100 to 200, and charting brings extra exposure.
• It most likely bought ads in these countries (inferred): it charted only in low-cost markets, its curve is step-shaped, and ad monetization offsets part of the install cost. The developer also built a small tool that auto-prices by each country's purchasing power, so he understands the emerging-market playbook.
• The website catches search traffic. scaleforgrams.glipo.org is written as a Q&A page, with titles like "Can I use my iPhone as a weighing scale?" and "Is there a free digital scale app for iPhone?", questions people actually search. The web version also offers one free trial, plus one more after logging in.

It has no content channels at all. Searching TikTok for "scale for grams" and "timbangan ai" (Indonesian for "AI scale") only returns product videos selling physical scales. Nobody mentions it on X, and there is no media coverage.

Why Android

"How to weigh something with your phone" is a very old need. In earlier years it was met by fake scales and prank apps that let you place an object on the screen and then show an arbitrary number. Starting in 2025, it was replaced by "take a photo and ask a large model."

On iOS this path is already crowded. There are over a hundred apps with both scale and gram / weigh in their names, and from 2025 May through 2026 October a new one appeared almost every week, most with single-digit ratings. The leader is another Turkish company's Scale For Grams: Digital Scale from AVLABS (launched 2025-05-24, about 3,500 ratings in the US store, translated into multiple languages including Italian, Spanish, and German, and charted in 14 countries). There's also an independent site, scaleforgrams.com, with web plus iOS, priced at $6.99/week or $49.99 one-time, running SEO comparison pages.

Android is much emptier:

In short: the same idea and the same search term is a red ocean on iOS, while on Android, apart from Primo Scale, almost everything sits at the 1,000-download level. During the weeks it gained traction, it was the only app in this category visible on the Android category chart.

That 1M-download Product Scanner also tells us two things: this keyword really does bring in massive downloads, and users who discover "it's not actually a scale" react hard, to the point it got renamed. Glipo's heavy disclaimers and constant use of "estimate" are probably there to dodge Google Play's review of "misleading functionality" (inference).

Where it falls short

It's inaccurate. A photo can't tell material or whether an object is hollow or solid, which is how a ring ends up off by six times. Negative reviews almost all say this, while positive ones are mostly just "I like this app." Users who estimate once, see it's wrong, and delete are probably common, and there's no retention data (inference).

Anyone can build it. ChatGPT or Gemini can take a photo and answer "how much does this weigh" about as well, and calorie apps like Cal AI already estimate food grams from photos. They don't own the "scale" search term, so they haven't taken this audience yet.

No moat. The product is a prompt plus a result card. One person could build it in a week, and iOS history in this category already proves that.

If it were you

C: worth learning from, not worth building an Android "AI phone scale" for. It shows that the "photo + AI" category, crowded on iOS, can be moved to Android and cheap countries, with the search term copied verbatim into the title, and a solo dev can still break into a category chart. But it only has 30-odd thousand downloads, never hit the Top Grossing chart, and on Android there are a dozen same-named apps plus Primo Scale with 100K downloads, and the product is inherently inaccurate.

Good for: indie devs who know store keywords, are willing to spend small amounts on acquisition tests in countries like Indonesia, Thailand, Mexico, and Brazil, and can quickly ship several "photo + LLM" mini tools.

Not good for: anyone wanting to build a genuinely useful weighing product (the accuracy ceiling of photo-based weight estimation is fixed, and no matter how well it's built, "six times off" reviews will hit it); anyone focused only on the US market who won't spend on ads; anyone who can't accept the math of "ads + subscriptions, with very little value per user."

How to do it: don't copy the scale, copy the method. Look on iOS for "photo + AI" needs that already have dozens of same-named clones and several thousand combined ratings, but whose Android top ten is filled with 1,000-downloaded small apps, such as counting items, measuring sizes, identifying objects, or estimating prices (figure out for yourself which category has gaps). Put the biggest search term directly in the title, pick a category with a low chart-entry bar, launch subscriptions and ads together, and write thorough disclaimers.

How to get users: run Google App ads in 2–3 low-cost countries, starting at a few dozen dollars a day, to see whether it can push into the top 100 of the category chart; build a website with several "Can I use my phone as a ___?"-style Q&A pages to catch search. Don't count on TikTok, since these tools have no content to film.

When: before it's too late: on Android, these gaps are filled one by one. A dozen people are already doing scales, so for other categories, the earlier the better. The window closes whenever Google tightens review of apps that "pretend to measure."

Signs you've misjudged it: ① after one or two weeks of paid acquisition, the cost per install exceeds what ads plus subscriptions can recover, and it can't crack the top 200 of the category chart; ② "inaccurate" or "scam" dominates the negative reviews and the rating drops below 3.5 stars, meaning the need can't retain users; ③ conversely, if Glipo stopped doing anything and Play installs kept climbing steadily and it started showing up on Top Grossing, that would mean this segment is bigger than it looks, and weighing itself is still worth building.