Cal AI revenue: the teen-built calorie app MyFitnessPal bought

Calorie counting has a famous failure mode: people start enthusiastically, log every meal for a week, and quit because typing in food is tedious. Cal AI attacked that single step. Take a photo of your plate, and the app estimates the nutrition. Two high-school students built it; by the time MyFitnessPal, the incumbent in the category, announced it had acquired them in March 2026, Cal AI’s trailing-12-month sales were over $40M (verified, MyFitnessPal press release).

Why did the founders build Cal AI?

Logging every meal by hand is the reason most people quit calorie counting. Vision models made it possible to estimate a meal from a single photo, removing most of the effort. The insight is not that people want to track calories; many already do. It is that the most painful step in an existing habit could now be automated.

Yadegari, who had already built and sold a small web business as a teenager, started Cal AI with a schoolmate, Henry Langmack, while both were still in high school. Forbes later described him as someone who kept launching apps until one worked (Forbes).

How does Cal AI work?

A user photographs a meal. The native mobile app sends the photo to vision-capable AI models and combines the result with nutrition data to estimate calories and macronutrients. The app then logs it, so the habit continues without manual entry. Access is through in-app subscriptions.

How was Cal AI built?

The team shipped an iPhone app first. The architecture is a native mobile app talking to third-party vision models plus a nutrition database, which meant two students could ship it without training their own models. The team stayed small: seven employees plus contractors at acquisition (media-reported, TechCrunch).

How did Cal AI market and sell the app?

Marketing leaned heavily on influencer and creator content on TikTok and Instagram, paid user acquisition, and tight paywall and onboarding experiments. In consumer subscription apps, the onboarding flow and paywall are where revenue is won or lost, so constant experimentation there is as important as the AI feature.

Pricing follows the usual model for consumer mobile apps: a free download with in-app subscriptions. Exact prices are not cited here.

What challenges did Cal AI face?

The biggest public challenge was the gap between projections and revenue. A widely repeated headline said Cal AI made $35M in a year. The founder-reported picture in May 2025 was different: about $10M had been collected year-to-date, and $35M was a forward 12-month projection (founder-reported, Latka). Readers should keep the two numbers apart.

A second challenge is structural: a consumer app that depends on creators and paid acquisition must keep acquisition costs below subscriber value, in a category where a giant incumbent already exists.

How much revenue did Cal AI make before the MyFitnessPal deal?

Cal AI’s monthly revenue is not disclosed, and its total revenue is not disclosed. The best figure is MyFitnessPal’s own: trailing-12-month sales of over $40M at acquisition (verified, MyFitnessPal press release). Forbes reported a $50M figure, which is a projection for 2026, not current revenue (media-reported, Forbes).

The timeline: the app launched in 2024; in May 2025 about $10M had been collected year-to-date; the acquisition by MyFitnessPal closed in December 2025 and was announced in March 2026, with terms undisclosed (verified, MyFitnessPal press release).

Why did Cal AI work?

It removed the most painful step in an existing habit. Customers did not need to be convinced that calorie tracking matters; they needed it to stop being tedious.

It also treated distribution and paywall design as core product work. Consumer AI apps win on distribution and paywall design as much as on the model. Umax shows the same creator-driven social growth in another consumer category, and Suno shows consumer AI subscriptions at much larger scale.

Finally, growing fast inside a category with a big incumbent creates an exit path. Incumbents buy fast-growing challengers.

What can founders learn from Cal AI?

Remove the most painful step in an existing habit. The best consumer AI ideas often shrink a familiar task to one tap.

Consumer AI apps win on distribution and paywall design. Budget as much effort for onboarding experiments and creator partnerships as for the AI.

Incumbents buy fast-growing challengers. If you build in a category with a dominant player, that player is a natural buyer.

And keep projections and revenue separate in public. The $35M headline was a forecast; the over-$40M figure is what the acquirer reported.