❓ Problem

Calorie tracking is the most abandoned habit in consumer software. Roughly 70% of users quit within two weeks, and the strongest predictor of whether someone is still logging six months later is not motivation or discipline. It is seconds per meal.

The industry already solved the slow part. Cal AI turned photo-in, macros-out into $30m of annual revenue and 15 million downloads in under two years, then sold to MyFitnessPal before its founder turned 20. Food recognition is a commodity now. Every serious tracker has it.

So the friction moved. What is left is everything wrapped around the recognition: find the app, download it, make an account, sit through an onboarding quiz, then remember it exists three days later and open a fourth icon at the dinner table while your food goes cold. Each of those is a small tax, and consumer habits die from small taxes.

And the incumbents feel it. MyFitnessPal has 220 million registered users and more than 30 million monthly actives, but its revenue slipped 5.7% to $310m in 2025. Two hundred and twenty million people signed up. Thirty million still open it. That gap is the entire opportunity, and it is a distribution problem dressed up as a product problem.

✅ Solution

A calorie tracker with no app, no account and no home screen icon. You text a photo of your lunch to a number and it texts back the macros. That is the whole product surface.

• The thread is the food diary. Already open, already on your phone, already the app you check fifty times a day. Scroll up and there is last Tuesday's dinner, with the photo attached. No export, no sync, no separate history screen.
• Onboarding is one message. No download, no account creation, no password. Text the number, answer three questions about your goal, start logging. Time from "I should track this" to first logged meal is under a minute, versus a download plus a signup flow plus a paywall.
• It texts you first. This is the retention mechanic nobody has. A message at 8pm asking how dinner went reads as a normal thing to receive. A push notification from an app you last opened nine days ago reads as an app begging. Same intent, completely different psychology, and it is the only lever in this category that meaningfully moves the two-week cliff.
• No app also means no App Store tax. Cal AI charges $29.99 a year and keeps roughly $21 of it. Bill $9.99 a month on the web and you keep about $116 on annual billing. Same customer, same job, several times the revenue per head.
• Then sell the seat, not just the subscription. Coaches, dietitians and GLP-1 clinics buy a dashboard that shows every client's log in real time. The client does nothing new. They just text, same as before.

The wedge is deliberately narrow: one number, one job, one thread. The platform underneath (a proactive nutrition agent with a professional dashboard on top) expands into the whole accountability layer.

📊 Key Numbers

Market size

• The global diet and nutrition apps market is worth around $2.8 billion in 2026, heading to $4.6 billion by 2030 at a 13.4% CAGR. North America is the biggest slice at 36.4%.
• MyFitnessPal alone does $310m a year off 30 million monthly actives, and its premium tier runs at $79.99 a year. Cal AI added $30m on top in two years from a standing start, which is the useful datapoint: this category still rewards a new entrant with a friction advantage.
• The serviceable slice is US and UK smartphone owners who have tried to track food and given up, plus the GLP-1 cohort. Gallup now has 11% of US adults currently taking a GLP-1 for weight loss, around 40 million people, up from 3% in 2024. Every one of them has been told to hit a protein target.