Property 01 of 50

foodphoto.ai

AI food photography for restaurants — and the only property in this book with confirmed revenue.

What this property was

FoodPhoto takes the phone photo a restaurant owner snaps of their burrito under bad fluorescent light and turns it into something that looks like it came out of a studio shoot — lit, styled, color-graded, ready for a delivery app listing or an Instagram grid. Under the hood it's a Next.js frontend talking to a Python FastAPI backend, with Postgres and Redis behind it and a smart router that picks the best image model for each request from a handful of providers. You buy credits through Stripe, you upload, you download.…

Evidence recorded in the manuscript

This is the one property where I can say it plainly: real strangers have paid real money, repeatedly. I won't dress it up with a vanity revenue figure, but FoodPhoto is the confirmed cash engine of the portfolio — which is exactly why its failures hurt the most. One morning a single Python import error crashed the backend and paid orders failed silently for most of a day before I caught it.…

The lesson recorded after launch

The property that makes money is the property where sloppiness becomes expensive. On a parked experiment, a broken checkout is trivia; on FoodPhoto it's an outage with a dollar cost and an angry customer attached. So this is where I learned the operational discipline the rest of the portfolio borrows: one builder per repo, snapshots before builds, verify the money path end-to-end after every change, and never trust a function that returns success without checking what it actually did. Earn revenue first; the revenue will then teach you everything else.

Related property case files