Snap a plate, get the calories: a Brazilian marketing firm localized Cal AI into Portuguese and hit #1 in the fitness free chart within three weeks

Americans have already been taught by Cal AI that a photo of a plate tells you how many calories it has, but Brazil never had a local version in Portuguese that understands Brazilian food. In Oct 8, a company in Maringá, Brazil, named "Marketing Affiliates", launched BeFitWay, wrapping it in a friends-group weekly leaderboard and fitness feed, and used ads, creators, and pay-per-view clip editors together to build volume. By Nov 5 it was #1 on Brazil's App Store Health & Fitness free chart and #9 on Top Grossing. Below: what it is, how it took off, where the money comes from, and whether you can do the same.

A "false advertising" one-star review

In Oct 29, a Brazilian user gave BeFitWay 2 stars, titled "Propaganda enganosa" (misleading ad), with one line in the body: "The ad says it's free, and after installing you can only use it if you pay", i.e. the ad says free, but after install only paying unlocks it.

This review reveals two things: ads were actually running and pitched it as "free"; and installing led straight to a paywall. The next day it added 74 ratings in the Brazil store, up from 9 the day before; from 10-30 to 11-01 it added about 170 ratings over three days (Brazil App Store rating count). The same week, the Android version on Google Play added about 7k installs in 11 days, for a total of 19,281 (Google Play, 11-06). A concentrated push most likely opened in those few days (inference).

What it is

Point your phone at a plate and AI estimates calories, protein, carbs, and fat; adjust the portion and save it to your food diary. The official tagline is "Conte calorias com uma única foto" (count calories with a single photo); the clip-contest brief is blunter: no scale needed, one snap tells you how heavy and how many calories the plate is.

The rest is standard: barcode scanning, fasting timer, water reminders, weight curve, 7 days of meal plans and shopping lists, and Apple Health and Strava integration. What sets it apart from Cal AI is that it layers in a social feature:

• Friend groups and weekly leaderboards: private groups ranked by weekly check-ins. This is basically GymRats' playbook, which is hugely popular in Brazil (GymRats has 148k ratings in the Brazil store).
• 24-hour feed: post a photo of your workout or meals every day. Posting earns "BeFitWay Coins", which you redeem for premium and discounts.
• Conexões Fit: added in 11-03's 1.0.6 update, a 18+ matching chat, "find people with the same goals, match and chat" (App Store release notes).

Portuguese only. App Store rating 4.77 stars from 260 ratings (11-06).

Photo calorie tracking is already crowded. What gets it in?

Photo calorie counting has no real moat on its own. It's most likely a multimodal LLM call (its privacy policy says data goes to a "natural language processing model" provider). What sets BeFitWay apart is three things: Portuguese-first localization, a bundle of "calorie tracking + GymRats-style group challenges", and a marketing agency running distribution.

Was social the reason it took off? Probably not the main one. The feed was already live in October. 10-25 A five-star review said "As redes sociais são demais, dão um gás a mais!" (the social stuff is awesome, it gives you a boost), but dating only came after the volume took off. That surge looks more like it was driven by paid acquisition.

Where users come from

Paid ads. Section 3(b) of the privacy policy says it integrates Firebase and Meta SDKs to "measure the effectiveness of our ads on Google, Instagram and Facebook". The events sent back are first open, sign-up, trial start and purchase. The opening negative review is direct evidence of ads. I couldn't find how much was spent or what the creatives looked like.

Influencers. The official Instagram @befitway has only 1,349 followers, but it has a story highlight called "Famosos" (celebrities). The website has a dedicated influencer link, befitway.com.br/fitness, which jumps straight to Google Play, with utm_medium=influencer in the URL. One link per channel means you can see exactly how many people each influencer sends. I couldn't identify which influencers.

Pay-per-view clippers. On Cut.Pro (a Brazilian clipping bounty platform) there's a "Campeonato de Cortes BeFitWay": 11-02 to 12-02, R$10,000 total prize pool, R$1 per 1,000 views, with a minimum of 3,000 views per clip and a cap of R$300. The rules are detailed: the video must show the "photo your plate, get weight and calories" moment, tag @befitway, the influencer link goes in the bio, bot-views disqualify, and the brand can repost your clip. 11-06,54 clippers submitted 117 approved videos, totaling 14k views, for only R$3 paid out (Cut.Pro, 11-06). This tactic is new. It didn't drive October's spike, but it's feeding future growth.

Store search. The title is literally "Contador de Calorias" (calorie counter), the biggest search term for this category in Brazil.

Ratings. Some look unnatural: "Muito prático, com uma foto eu tenho todas calorias, valeu!!" appears word for word in both 10-17 and 11-01. There are also several five-star reviews that just say "5", "Bom" or "Shontelle". Whether reviews were incentivized is unverified.

The official TikTok has 132 followers and 4 posts. Nobody discusses it on X. Outside Brazil, almost nobody knows this app.

Where the money comes from

3-day free trial, then a hard paywall. Brazil App Store pricing (11-06): monthly R$29.90, "special" monthly R$19.90 and yearly R$82.90, plus two more tiers at R$44.90 and R$149.90 (probably quarterly and annual); coins come in packs of 500 for R$4.90 up to 8,000 for R$39.90. The website's terms include refund rules for "direct purchases on the website", which means there's web checkout that sidesteps Apple's commission.

It climbed from #139 to #9 on Brazil's Health & Fitness Top Grossing chart in a week (10-28 to 11-05), and 11-05 it also hit #180 on the overall Top Grossing chart, ahead of BitePal and MyFitnessPal. Revenue isn't public, but going by that rank, it's likely a few hundred dollars a day (est.).

If it were you

B. What it proves isn't a single product but a playbook: take an AI tool already proven in the US, localize it for a large non-English market, and seed it through the local affiliate-marketing scene. It reached the top 10 of its category's Top Grossing within three weeks. But it has no moat of its own. Photo calorie counting is something anyone can build, Brazil already has Foodvisor, FoodPilot and Monju ahead, and its growth depends on paid ads, which is harder than writing the code.

Good fit for: small teams that can run Meta ads and do the math on trial-to-paid ROI; or people in a large non-English market (Indonesia, Turkey, Vietnam, Mexico, Poland, that kind of place) who know local influencers and clippers.

Bad fit for: indie devs who only want to build and won't spend on acquisition, since this category doesn't take off on organic traffic; anyone who wants to go head-to-head in the English market (Cal AI's home turf); anyone who cares about reputation, because "ads say free, but you pay after install" complaints are a side effect of this playbook.

How to build it: pick the market first. Check where Cal AI ranks in that country's Health & Fitness Top Grossing. The lower it sits, the more likely nobody has localized well. The core product is photo-to-calories plus recognition of local food (in Brazil that's padaria pastries and churrasco; train on whatever your local food is), plus a friend-group weekly leaderboard to give people a reason to come back daily. 3-day trial, hard paywall, and web checkout on the side.

How to acquire users: copy its three-part stack. Run Meta ads that send trial and purchase events back, and adjust by ROI. Give each influencer a UTM-tagged link. Run a per-thousand-views contest on a local clipping bounty platform, with a rule that the "snap it, get the calories" moment must appear on screen. That's the best ad creative for this category, and you only pay for results.

Deadline: the window for localizing "AI photo calorie tracking" is being filled country by country (the Hong Kong team has already shipped a Portuguese version in Brazil). Whichever market gets taken first is one fewer opening. I expect few markets will still be open in six months (inference).

What would change the call: (1) if after December it drops out of Brazil's Health & Fitness top 50 Top Grossing, the October spike was paid-driven and the playbook's payback is worse than it looks; (2) if it holds the top 10 and starts launching in other countries, localization plus social really retained users, and it's worth upgrading to A; (3) if Cal AI or Foodvisor is already in the top 20 of your chosen market, there's no opening there, so pick another.

---

Data and sources: App Store Brazil Health & Fitness free and grossing chart positions, ratings count and reviews (10-17 to 11-06); iTunes lookup (Brazil store, 11-06, ratings, version, release notes); App Store Brazil listing pricing (11-06); Google Play installs and rank (11-06); befitway.com.br, befitway.com.br/fitness, and befitway.com privacy policy and terms of service; the Cut.Pro "Campeonato de Cortes BeFitWay" page (11-06); Instagram @befitway and TikTok @befitway (11-06); Brazilian company registration records. Where I wrote "estimate", "inference" or "unverified", that's what those are.