A savings app that splits your income into pockets the moment it arrives, and locks the important ones until they're due. I built the design system and the screen layer.
Your salary lands. By the 9th, it is gone. A few people manage this in a spreadsheet, which nobody enjoys. Most of us wing it.
The existing answer is a susu collector, who takes a cut whether you stick to your plan or not. We wanted the opposite arrangement: costs nothing unless you break your own promise.
That meant it couldn't be a budgeting tracker that watches money leave. It had to be the account the money actually lives in.
The design system: the theme, the component library, and the screen layer built on top of it. Pockets, balances, activity, onboarding, profile, and the authentication screens.
A preview harness that renders every component in isolation and against live data, so the system could be reviewed as a system rather than only ever seen inside a finished page.
The public landing pages.
One source of truth for type, colour and spacing, written first. On a three-person build with two other people moving fast, a shared theme is what stops the interface drifting into three different products.
The pocket row, the pocket card, the activity list, the empty and loading states, all built as a library first. Screens are then composed from pieces that already agree with each other, which is much faster than reconciling them afterwards.
Components are easy to judge in isolation and hard to judge buried three screens deep behind an authentication flow. The preview route made the whole system reviewable at once, including against real data.
It installs as an app on a phone and is designed for nothing else. Deciding that early meant no compromises made for a desktop layout nobody was going to use.
Live, and moving real money. It runs over a payment rail licensed by the Bank of Ghana, which holds and moves the funds; Prodasher provides the pockets, the automatic splits and the commitment locks.
Pockets can be funded by dialling a USSD code from a feature phone, with no internet connection and no app installed. That one detail is the difference between a product for people with good phones and a product for people.
This was a three-person build and I want to be exact about my part in it. I built the design system, the component layer, the screens and the landing pages. The backend, the payment integration and the security architecture were not mine. It is the only project here I didn't build alone, and the only one you can click into and use.