Loading page…
Snap a photo of your food, and AI tells you how many calories are in the meal
peak monthly revenue ≈$3M
A high schooler turned his own gym-life calorie-counting hassle into a product: 5M downloads in 8 months — the pain point was his own, and the demo takes 3 seconds.
Fewer bars = easier, cheaper, or faster for an AI-assisted solo builder. Editorial judgments based on the case details.
Zach started lifting in high school to get girls' attention. Fitness is "30% training, 70% diet," and he found calculating how many calories he ate every day too much of a hassle — so he pulled together a few friends and built an app: photograph your food, and AI automatically calculates the calories. The crew coded day and night in a San Francisco hacker house, and the app exploded the moment it launched: 5M downloads in 8 months. Note the pattern: not chasing a trend — eliminating a real hassle of his own.
Cal AI did not creep into the market. They shipped, and in Zach's words the response was immediately very positive, which is what convinced the founders to go all in. Everything after that is bought rather than stumbled into: he names influencer marketing as the company's special sauce and says the far majority of spending goes there. The product was built for that channel whether or not anyone planned it that way. Pointing a phone at a plate and getting calories and macros back is a three-second demonstration, so a creator's video is the ad and the proof in the same clip, with no setup and nothing to explain. The starting capital came from his previous project -- everything he had made from Totally Science, the unblocked gaming site he built as a high school freshman, with a co-founder doubling it. That is the real entry requirement here. Writing the app cost almost nothing because he and his CTO built it; the money only started moving at public launch, and it went to creators.
Revenue runs entirely through Apple. Subscriptions are billed by the App Store and even refunds are handled by Apple rather than by the company, and that detail sets the shape of the early months. Apple pays out roughly 45 days after a sale, and because the team was buying creator promotion faster than Apple was settling, they spent months putting money in without a proportional amount coming back out. It took about six months to get ahead of the payout schedule, and as Zach is careful to point out, they were very profitable that entire time. Profit and cash are separate problems, and this model front-loads the second one. What it grew into explains why the squeeze was worth funding: about $1.4 million a month after the App Store's cut, roughly $25 million a year net on a team of four, with a peak month close to $3 million. Copying it means holding a float that covers six weeks of ad spend before any of it returns.
Proves the ceiling of "small life pain + minimal app + influencer distribution" is in the tens of millions. Do websites/plugins first; save this model for when your pipeline is mature.
Paste into Claude Code / Codex and get a working version of this product end-to-end.
Two production-grade prompts per idea — a full build spec and an SEO growth plan — plus 10 tailored build specs a week. Members only.
Join the Unicorn Club — $5/moCancel anytime, in one clickEvery claim traces back to a listed source — open any of them and check the number yourself.
Data credibility: Cross-confirmed by two independent sources. All revenue figures are founder-reported or estimated by the video creator; none are audited.
Zach names the turning point himself, and it is not a growth tactic. He and his CTO moved to San Francisco for July, lived in a hacker house and effectively slept in the office, and that month is when they made their first real hires, built a team, and settled into a rhythm that carried the next year. The mechanical loop came afterwards, and it is a cash loop rather than a viral one: creator spend buys installs, installs become App Store subscriptions, and once the company was ahead of Apple's 45-day settlement it could fund the next wave of creators out of revenue rather than out of savings. That is how 5 million downloads stack up in eight months. Retention decides whether the loop holds, because paid growth stops the day the spending stops, and it is why the app teaches scanner technique in the tutorial: photograph a bowl with ingredients hidden underneath and the number is wrong, and a wrong number on day one ends the subscription.
The hard stretch here was financial rather than emotional. For roughly six months the company was paying creators faster than Apple was paying the company, so the bank balance looked like a business in trouble while the business itself was very profitable. What kept the judgement straight was watching the right number: the unit economics of each cohort, not the cash left at month end. When a dollar of creator spend reliably returns more than a dollar of subscription revenue, a widening cash gap is a financing problem rather than a verdict on the product. Zach's advice for the start of that period runs against caution -- stay away from over-testing early, move fast, launch fast, trust your instinct -- and his own pace was around 40 hours a week on Cal AI alongside high school, to the point of sitting through AP exams without taking them. The launch reaction is what bought that conviction, and it is the signal they went all in on.
The roundup video: $25M in net annual revenue, a team of 4. A live interview with Zach himself at an app-founder conference: "my best month was close to $3 million in revenue" and "I'm only eighteen." The two sources line up — among the most credible numbers in this entire dataset. He's now preparing to be CEO while attending college "to experience a normal person's life."
At the conference, multiple consumer-app operators making $1M+/month all name-checked the same thing: influencer marketing is the most undervalued channel right now — "done right, it's extremely profitable; use it as much as you possibly can." (My take: Cal AI is the signature case of TikTok/IG fitness-influencer campaigns — the "snap a photo → calories appear" 3-second demo is born for short video; the demo IS the ad, and that's the root reason it could blow up on paid.) Zach's own advice for early-stage founders: "Early on, stay away from (over-)testing. You must move fast, launch fast, and trust your gut."
It proves the ceiling of "small life pain + minimal app + influencer distribution" is in the tens of millions of dollars. The transferable formula: find a daily behavior that's high-frequency + annoying + demoable in 3 seconds — snap a photo to identify an outfit, snap a photo to get a room-makeover plan, snap a photo to price secondhand items... "point the camera at X" has been widely copied; the winning move is finding a new X. For you: this takes iOS development + an ad budget + a feel for the North American content ecosystem — slot it into the phase after your websites/plugins are proven. Risk: stop the ad spend and growth stops; with weak retention it's just a leaky bucket.
That was one of the 50 breakdowns open to everyone. The Unicorn Club opens the other 1,285 at the same depth — plus pain-point filters and 10 build specs a week.
Join the Unicorn Club — $5/moCancel anytime, in one click