Tomer Cnaan did not invent anything. That is precisely the point of his story.
He was listening to a podcast interview with Zach Yadegari, the teenage founder of Cal AI — the American app that estimates calories from a photo of your meal and was, at the time, doing over a million dollars a month in the US. Tomer's thought was not "I should compete with that." It was: "This does not exist in Israel yet."
So he built CalBuddy: the same core interaction — photograph your food, get calories and macros back — but Hebrew-first, tuned to Israeli cuisine, Israeli language, and Israeli users. According to Tomer's own account in a Starter Story interview (founder-reported figures, not audited), the numbers went like this:
- June 2025: launch. First month: about $1,028.
- Month four: roughly $20,000 a month, driven by paid ads. His edge was boring and decisive — Israeli ad inventory was far cheaper than US ad inventory, and he studied both the winning American calorie-app creatives and high-performing local fitness content before spending.
- April 2026: over $81,000 in a single month — an annual run rate around a million dollars, inside a year.
The jump from $20K to $80K did not come from more ads. Tomer brought in two well-known local fitness faces as partners on a revenue-share deal — reported as a 50/50 split. They supplied the two things a cloned product cannot copy: local trust and local distribution.
What he actually copied — and what he didn't
The feature was copied, and the feature is the least valuable part. Photo-in, calories-out is a commodity interaction now; any competent builder with an AI coding tool (Tomer built with Cursor) can reproduce it in weeks.
What he built himself was the commercial system around it:
- A real local food model. An AI tracker that misreads the dishes people actually eat is useless. Local cuisine was not a translation task; it was the product.
- Cheaper attention. The same ad playbook that costs a fortune in the US ran at a fraction of the price in a smaller market.
- Distribution with skin in the game. Revenue-share partners promote like owners, because they are.
Meanwhile the original kept compounding: Cal AI itself went on to be acquired by MyFitnessPal, announced in March 2026, after passing a reported $30–50 million in annual revenue. The pioneer got the exit. The localizer got a million-dollar run rate in a country of ten million people. Both outcomes came from the same interaction.
The transferable lesson
Most builders ask: "What new thing can I make?" Tomer asked a better question: "What already works somewhere else, that my market does not have yet?"
An idea that is old in one market can be brand new in another. Demand proven in one geography does not lock the rest of the map. The moat is not the idea and not even the code — it is how deeply you localize: language, food, media prices, and faces people trust.
You do not always need an original idea. Sometimes you need a passport for one.
— Terry, Airgrix
Top comments (0)