Picture this:

Someone's making a tasty, but hefty ass sandwich on your FYP

Cottage cheese, avocado, boiled eggs.

Nothing special.

But then halfway through, she mentions an app, drops a code in the caption, and goes back to munching…

You keep scrolling… but what you don’t realize is that behind this fifteen-second clip is basically the entire marketing budget behind a consumer app that just got bought by MyFitnessPal.

And yes, we’re talking about the legendary rise of Cal AI.

We almost skipped this case study, honestly.

A photo-of-your-food calorie app doesn't sound like a story worth your five minutes, and not to mention:

You’ve probably heard this story a HUNDRED times…

But it’s starts to raise eyebrows once you look at how it actually got bought.

Why?

Two teenagers built it.

Zach Yadegari and Henry Langmack, who met at coding camp when they were just ten years old.

$30M+ in revenue, & 15M+ downloads - all in under two years of grinding.

I couldn’t even get a single day one to lock in on Online Business with me… so yeah, I’m just a *little* bit jealous right now.

But from every source I’ve on their “zero-to-hero” story, I start to see this never-ending pattern:

They only cover their surface win that made the headlines.

So let me be the first to actually go a little deeper into it, using data no other source could POSSIBLY provide on this.

And I’ll start with what *should* be a common belief… but if your brain still lives in pre-lockdown era?

None of their initial success came from running ads.

From just the ripe age of 7, Yadegari had already shipped a few apps before this one:

Normal teenage-kid stuff that’d probably wouldn’t get you popular with the ladies at lunchtime 🤷🏽‍♂️.

(Correct me if I’m wrong 😹.)

But it’s the kind of thing that teaches you years of experience in the game without ever actually studying or working to gain it.

Calorie tracking wasn't exactly “wide open” either.

MyFitnessPal had run the category for a decade - Foodvisor and MacroFactor were fighting over what was left.

Cal AI's whole pitch: point your phone at your food, let it guess the calories.

Nothing revolutionary. Just faster.

And in it’s earliest stages?

Brother… it was nothing but a GPT-wrapper.

So how do you beat a decade-old incumbent with an idea that small, and all odds stacked against you?

They utilized an opportunity none of the old heads could see at the time.

UGC.

But remember a second ago when I said he developed a website blocker - yeah the one the *sounded* like another throwaway that got a couple clicks?

Well Zach sold it for six-figures to a company called FreezeNova when he was around 16. And that exit cash was the seed money behind making Cal AI more than just a huge success.

Not an investment round with corporate suits… not savings from a 9-5, stacked on top of a temp summer job.

But one bet that paid off, put straight into the next one.

And with just around $100,000 liquid cash sitting impatiently in his junior checking account, Cal AI didn’t wait around for a video to just magically go viral.

He took that risk.

He paid for over 250 UGC creators, on retainer, every month to post “watch me scan my food“ ads into their normal daily-routine content.

Three seconds, one mention of his app, then back to their lives on a screen.

$30,000 in the first month.

Over $100,000 in the second.

And over $1 million six months later…

Once Cal AI started to take off, their playbook widened:

Bigger creators with larger audiences, A six-figure placement with MrBeast to make their app feel like a real brand.

Then finally, in-house paid ads across TikTok, Instagram, and Facebook, which by the time MyFitnessPal came calling were running past $1M/mo against a new high of $5.7M in MRR.

And somewhere in that same stretch, the retainers turned into something cheaper to run: an affiliate code, one per creator, no monthly fee, just a cut of whatever came through their link.

So basically, CPMs.

Picture what that actually means day to day.

Every creator running a code is its own mini business: its own audience, its own cost to produce (a phone and twenty minutes, nothing more), its own conversion rate Cal AI can watch in real time, same as a real business owner checking the register at close.

The ones pulling downloads get more attention.

Maybe a bigger deal next time, maybe Cal AI just goes and finds five more creators who look and sound like them.

The ones that don't convert get nothing dramatic.

No firing, no phone call. The code just stops getting renewed, and outside of Cal AI's own dashboard, nobody ever finds out that creator was even part of the plan.

So as you can see, if one code flops, you've lost nothing.

One code works → you copy it (New creator, new code, same watch.)

NOBODY has to go viral.

The system just had to work often enough, and Cal AI could always tell when it did.

Now here’s how they scaled it.

Three creators or three hundred, the game doesn’t change: post, watch the code, keep what works, drop what doesn't.

And it’s not like the retainers and the ads didn’t contribute to the $30 million mark.

But this UGC layer is just the cheapest one to run, and the easiest one to keep scaling without unhealthily increasing spend.

(Somewhere, a marketing agency is still building a $10K strategy deck to explain what one random UGC creator just did for free.)

And hundreds, if not, THOUSANDS of apps, brands and AI tools have caught on to this playbook.

Just look at how many competitors I found copying his viral UGC playbook, when I searched for videos on his app using this AI tool.

Now I don’t even have TikTok installed…

So a question you probably have for me right now is… how on Earth do I know all of this, and how am I getting all this data to prove it?

WHENEVER I want do do content research, I just use Virlo.

It tells me exactly what’s going viral in my niche, by launching content marketing agents that:

And it’s completely free to start using right now → virlo.ai

Just to prove it to you, here’s a few Cal AI competitors I found in less than 30 seconds using Virlo:

if you're not following yet: