EN
Back to the archive

The archive · Consumer Apps · Product decision · 2024–2026

Cal AI's photo-logging bet: 15M downloads, $30M revenue, then bought by MyFitnessPal

Teens bet AI food-photo recognition could beat manual logging; Cal AI hit 15M+ downloads and $30M+ revenue, then MyFitnessPal acquired it.

Cal AI

The betThat AI photo recognition is accurate enough that users pay for speed — snap a photo, calories logged — so a seven-person teen team can outrank incumbents.Live

What the business is

Cal AI is an AI calorie-counter app that logs calories and macros from a photo of your food, built by teenage founders.

How it started

Zach Yadegari, then 17, had already sold his unblocked-games site Totally Science for $100,000 at age 16 and met app developer Blake Anderson on X. The idea came from his own goal of gaining weight at the gym and his hatred of typing meals into a tracker; he and high-school friend Henry Langmack built the prototype in a San Francisco hacker house.

What happened

Launched May 2024. By March 2025 the founders claimed over 5 million downloads in eight months, more than $2 million revenue in a single month and 30%+ retention, backed by a 4.8-star App Store rating on 66,000 reviews. MyFitnessPal's CEO said his team, which tracks roughly 70 competitors on Sensor Tower, spotted Cal AI climbing app-store charts in early 2025; talks lasted almost a year and closed in December 2025.

How it ended up

Acquired by MyFitnessPal, announced 2026-03-02; terms undisclosed. The seven-person team was retained, the app stays independent, and Cal AI was connected to MFP's database of 20 million foods. Co-founder Yadegari kept running the app while starting college.

Background

Cal AI is an AI calorie-counter that logs calories and macros from a food photo, launched in May 2024 by high-school friends Zach Yadegari and Henry Langmack together with app developer Blake Anderson. Yadegari had already sold an unblocked-games site, Totally Science, for $100,000 at 16; the idea for Cal AI came from his own bulking goal and his refusal to hand-type every meal.

The bet was that large image models plus retrieval over open food databases could make photo logging accurate enough (the founders claimed 90%) that users would pay for speed over the manual entry incumbents required. Growth ran on influencer and UGC marketing: by March 2025 the founders claimed 5M downloads in eight months, over $2M revenue in one month and 30% retention, with a 4.8-star App Store rating on 66,000 reviews.

The rise was visible on app-store charts: MyFitnessPal's CEO says his team watches about 70 competitors on Sensor Tower and first noticed Cal AI early in 2025. After almost a year of talks, MyFitnessPal completed the acquisition in December 2025 and announced it on March 2, 2026. The seven-person team stayed, Cal AI remains an independent app, and it was integrated with MFP's database of 20 million foods.

What has to be true

  • Photo-based logging removes the friction that kills diet apps — one tap instead of searching a food database — which let a seven-person teen team compete with a category giant.
  • Building on OpenAI/Anthropic models with RAG over open food data let the founders ship in weeks without owning a proprietary nutrition dataset.
  • App-store charts made the startup's growth a public signal: the eventual acquirer says it spotted Cal AI on Sensor Tower early, turning distribution data into acquisition interest.
  • Staying lean (seven employees, Sunday-night stand-ups) kept costs low enough that a ~$30M revenue run rate made acquisition an obvious outcome.

What can be applied

When the incumbent owns database accuracy, attack the experience gap: Cal AI won on speed and simplicity, and public app-store chart data made its rise visible to the acquirer long before any deal.

Aftermath

As of September 2026 Cal AI operates as an independent unit inside MyFitnessPal, with co-founder Zach Yadegari running it while attending college. The app was given access to MFP's 20-million-food nutrition database, and MyFitnessPal's CEO says there are no plans to fold it into the main app, arguing the two serve different user needs — Cal AI for speed, MFP for fine-grained accuracy. Deal terms were not disclosed.

Sources

spotted an error? The archive wants to know.

Your turn

You just read one. Describe what you are building, and see who is betting on the same thing.

Free account · 3 free questions · no card

Related cases