Charge in WhatsApp first, then ship an app: Indonesia's photo calorie tracker Gizi AI hit the top 10 Top Grossing within a month and a bit of launch

Indonesia's Gizi AI tells you the calories in a photo of your meal, plus an AI coach that nags you daily. It only reached iOS in 19 Sep 2026, with a little over a hundred ratings, yet it ranks above every calorie app on Indonesia's health and fitness top-grossing chart. The reason isn't the app. Before it existed, it had been taking payments on WhatsApp for over a year and running a team of TikTok creators selling discount codes. Anyone can build the product. What's worth copying is the order.

One video with 1.12M views, and no app yet

In 27 Jul 2026, a TikTok account called @giziwithalex posted a video titled "cara jadi cowo 10/10" (how to become a 10 guy), tagged #giziai #gym. By Nov 6 it had 1.12M views and 140k likes. The account had just over two thousand followers, and its bio read "25% OFF CODE ALEX".

Gizi AI had no app then. The video sent people to a WhatsApp number: you send a photo of your meal, and it replies with calories and nutrients, logging everything daily.

7 weeks later, in Sep 19, the same account posted again: "Finally GIZI AI sekarang udah ada aplikasinya" (Gizi AI finally has an app). It only got 8,791 views, but that didn't matter, since most people watching were already paying users. People paid inside the app on day one.

What it is

A photo calorie tracker plus AI nutrition coach app on iOS and Android, mainly in Indonesia, with Thailand second. The App Store page pitches:

• Photo-based calorie and macro estimates, broken down by dish component, with explicit support for Southeast Asian food like nasi lemak, laksa, chicken rice and mamak dishes
• Two coach personas to pick from: the strict "Coach Gizi" and the laid-back "Gisel". They check in daily to ask what you ate, and you can ask them how to order off a menu
• The app, WhatsApp and LINE all log into one diary

Nutrition data comes from Indonesia's Ministry of Health food composition table, plus brand-published data (momsmoney.kontan.co.id, 2026-04-06).

It was built by two Indonesians: CEO Faaiz Falih and co-founder Dafau Guciano. In that interview Dafau says he's from Pekanbaru in Riau, "where everything is curry and chili sauce, delicious but not healthy". It launched in July 2025, took no funding and ran on subscription revenue alone. The developer account is listed under Singapore company Nalar Intelligence, incorporated in February 2026.

How big it is now

Number 8 on the top-grossing chart with only a hundred-plus ratings says it all: far more people pay than leave reviews, and payments are coming in faster than installs. On the same Indonesian top-grossing chart, InOut is #9, Calz #13, MyFitnessPal #15, fatsecret #17, and the US hit Cal AI #98 (App Store Indonesia, 11-06). Ahead of it are Strava, Flo, Runna, Huawei Health and Fitbit.

The store listing says "20,000+ people started before you". Not verified.

Why this one: three things were ready before the app launched

First, paying legacy users. Gizi AI never offered a free tier. Dafau's words in the interview: "No free version, everything is paid. If people won't pay from the start, the product isn't valuable enough." The main entry point was WhatsApp, because "everyone in Indonesia uses it, even parents who aren't good with phones can use it directly without learning a new app." App reviews show people moving over: "Used Gizi AI on WhatsApp for a long time, the app is much more convenient" (Indonesia store, 9-29).

Second, a creator team on revenue share. The website has a creator signup page at /affiliate/signup. Each creator gets a discount code and a referral link, followers get 25% off, and creators earn a commission on sales. Searching "gizi ai" on TikTok turns up a string of code-carrying accounts: @giziwithalex (code ALEX), @kalofy_ (260 followers, top single video 57k views), @fitwith_jason, @rzkyptrmaa, @marvingiziai. Their content is very consistent: hungry late at night and wanting fried rice, desserts with no nutrition label, whether gorengan will make you fat, ending with "just snap a photo and you'll know".

At least some of these are company people. The two behind @desi.healthyera and @fitbybim list themselves on LinkedIn as Gizi AI staff and content creators, and the endorsement on the official homepage, "Coach Bima Wibowo", comes from the latter.

Thailand uses outside small creators: muay thai, fitness and eating-show accounts tag @Gizi AI Thailand in their videos and post a 25% discount code in the comments. The official Thai account has only 26 followers.

Third, the co-founder is a fat-loss creator himself. Dafau Guciano has 49k TikTok followers and 2.16M total likes, posting under #leanislaw. In 9 Jan 2026 he posted "want to count calories? Use Gizi AI, and the link in my comments gets you a discount", which got 66k views. Later entrants can't get that.

Add the timing: on Indonesian TikTok, #defisitkalori (calorie deficit) and #trackkalori keep trending, while overseas AI calorie apps have barely localized for Indonesia. Cal AI ranking only #98 in Indonesia's top-grossing chart is the evidence.

Where the money comes from

Subscription, prepaid up front. The early-bird price on the Indonesia website is Rp99,000 for 1 month(s) (about $6, original price shown as Rp198,000), and Rp237,000 for 3 months (about $14). In-app purchases on Google Play range from $9.99–29.99. How that relates to the web prices is unverified.

The website FAQ compares it directly with MyFitnessPal and Lifesum: those have "few Indonesian dishes, premium over Rp200k a month", while Gizi AI is "7 times cheaper". At the early-bird monthly price, the real saving is closer to half, so the "7 times" figure probably assumes annual billing or another plan (inferred).

Revenue isn't public, and no founder has posted numbers. An app charging $6 a month and ranking in the top 10 of Indonesia's health and fitness top-grossing chart is probably not making huge money (inferred). But with no free tier and no ads, every user is a paying user.

Bad reviews cluster on two complaints: "I paid, and after a week it made me log in again and pay again, and I can't get my old account back" (Indonesia store, 10-26,1 stars); "AI chat for logging exercise doesn't work well" (10-24,3 stars). Some also say "the WhatsApp version is more accurate than the app" (10-04). Thailand reviews are almost all five-star "snap a photo and it analyzes everything, it's great".

Who it's competing with

Technically there's no moat. One user said it 'mistook banana blossom for salted fish.' In Indonesia it wins by getting there first, charging first and lining up creators first; in Thailand, against a local leader like Kalguroo, the same playbook only reached #41.

Can you follow it?

Building another one in Indonesia isn't worth it. Gizi AI already owns the 'local food + WhatsApp + cheap' positioning, has old users and a creator network, and copycats like Giziku AI, BARA and a batch of solo developers' 'Kalori AI' apps are already in.

Going to a different country is a real opening. Here are the highest-ranked calorie apps on the health & fitness Top Grossing charts across 11-06 countries' App Stores:

Markets like Vietnam, the Philippines and Egypt: the top of the Top Grossing charts are global universal apps with tens to hundreds of ratings, the entry bar is low, and not one player focuses on local food.

Note that each country uses a different chat app: Indonesia and Malaysia use WhatsApp, Thailand uses LINE, Vietnam uses Zalo, the Philippines uses Messenger. Gizi AI's approach works because 'you log it inside the chat app everyone already uses.' Changing countries means changing the entry point. Also, since 2026, Meta has restricted general-purpose AI assistants on the WhatsApp Business API; a bot that only logs food probably doesn't count as 'general-purpose,' but read the terms before you start (inference).

If it were you

One-line verdict: Worth trying, but don't follow into Indonesia. Pick a country whose Top Grossing charts don't yet have a local calorie app, and copy its order: collect money in the chat app first, then launch an App.

Who it's for: One or two people who know the language and food of a Southeast Asian, Middle Eastern or South Asian market, can find a dozen local fitness and weight-loss creators to negotiate revenue shares with, and can work with multimodal models and WhatsApp / Zalo / LINE bot APIs.

Who it's not for: Anyone who wants to build a single English universal version and fight Cal AI and friends with paid ads; anyone unwilling to build local nutrition data; anyone who wants to add another app in a market like Indonesia, Thailand or Saudi Arabia where a local leader already exists.

How to build it: In the first month, build only the chatbot: snap a photo, get calories back, and a scheduled daily prompt asking 'what did you eat today?' No free tier; charge from a few dollars a month. Build the nutrition base from the country's public food composition tables, then calibrate the 20 to 30 most-eaten local dishes one by one. Once you have a few hundred paying users and know how they use it daily, launch the App and move old users over; it earns from day one, and chart rank will rise before ratings do. Set up accounts properly from day one so you don't repeat the 'paid, switched logins, lost account' one-star reviews.

How to acquire users: Pay TikTok creators per conversion, give followers a 25% discount code, and pay creators a share of sales. Then run two or three of your own 'weight-loss diary' accounts posting daily about how many calories local late-night snacks, fried foods and desserts really have. Price against MyFitnessPal's local subscription so people see at a glance that you're cheaper.

When to move: Gizi AI is already expanding into Malaysia and Singapore, and big players like Cal AI will localize eventually. To claim a country, get the chat version running before the first half of 2027. Indonesia's experience: it took more than a year from chat version launch to the App charting.

What would prove you wrong: (1) In your chosen country, the chat version can't get a hundred or two paying users within three months, meaning people there won't pay to 'track calories' or the chat app entry point doesn't work; (2) Gizi AI itself enters markets like Malaysia or Vietnam and only reaches the #40 to #50 range, as it did in Thailand, showing its Indonesia success relied mostly on the founders and old users, not a playbook that travels; (3) big developers like Cal AI or Calz start localizing food and language by country and reach the top 10 on local Top Grossing charts, showing the gap is closing.

Data and sources

• App Store charts, ratings, daily new ratings and reviews for Indonesia, Thailand and other stores (2026-10-28 to 11-06); calorie apps on health & fitness Top Grossing charts by country (11-06).
• Install counts (10-29, 11-06) and in-app purchase ranges from the Google Play listing com.giziai.app.
• Interview on momsmoney.kontan.co.id (2026-04-06).
• giziai.com, giziai.com/id (FAQ, pricing), giziai.com/affiliate/signup.
• TikTok @giziwithalex profile videos (11-06), @kalofy_, @fitwith_jason, @dafauguciano, @giziai.th and others; Gizi AI's company page and employee pages on LinkedIn.
• giziku-app.com, usebara.com.
• Where it says "inferred" or "unverified," it's inference or not yet verified.