Five prototypes in one day
On day one, my co-founder tested five clickable prototypes. Her words replaced ours, and by the evening the MVP was cut down to five screens and eight journeys.
This is the app I am building with a co-founder who is a professional model. This page covers the method: how it went from an idea to a working app, with a brand, a filed trademark and a tested security base, in about seven working days. Then how that method became a guide for the next app.

My co-founder knows the work of a professional model from the inside. I bring product, brand and build. Claude researches, drafts and writes the code.
The product stays private until launch. What I can share is how we work, because that part is reusable.
Brand, trademark, design, accounts, legal and code, in about seven working days of my time.
With one rule set on day four and kept since: security comes first, and every protection is proven by a test.
The app was built one way. That way is now a guide and a starter kit for the next app.
Project rules, a decision log and the security rule go in before any work.
Product, brand, accounts and legal run side by side, in the order their dependencies allow.
Each batch is tried on a real phone and closes with a proof sheet.
Day one kit
Copied into every new project
Product
With the domain expert
In parallel
Legal and GDPR in the background
Build
About 80 pull requests
On day one, my co-founder tested five clickable prototypes. Her words replaced ours, and by the evening the MVP was cut down to five screens and eight journeys.
A dated log where each decision gives its reason and names the one it replaces. 180 decisions in the first twelve days, so any session can pick up without re-explaining.
D5825 SeptMVP cut to 5 screens and 8 journeysD7428 SeptRule no. 1: security firstD10229 SeptA real app on the stores replaces the web appD1524 OctChecks run only when their code changesEach batch closes with a proof sheet. A counter-test breaks a protection on purpose to check that the test catches it; about ten were closed without merging. The AI has no access to production.
Claude prepares and writes. I decide, create the accounts and press the final button in production. My co-founder brings the craft and tests the app.
On day twelve, a retrospective turned the project into a ten-phase guide and a starter kit.