BriefQR

The people who most need time tracking are the ones least able to use it. Cleaners, site crews, kitchen staff: high turnover, shared phones, no company email. Any system that opens with "create an account" has already lost. It also had to be honest, and a printed QR taped to a wall gets photographed once and then scanned from the pub.

And nobody replaces their payroll system to buy a time clock. Whatever came out the other end had to land in the tool the customer already pays for.

Operators have no accounts at all. Identity lives on the device side: a small ESP32 unit at the entrance displays a QR that rotates on a short interval, so a scan proves presence at a time and a place. A photo of yesterday's code is worth nothing. Where the stakes are lower, a printed static QR does the same job for free.

The worker's side is a PWA: any phone camera, nothing to install, no app store review to sit through. Managers get a separate admin app for workers, devices, locations and reports.

Export stays deliberately provider-neutral, with signed webhooks alongside a pull API and a Kimai connector written as the reference implementation. Adding the next payroll target means writing a connector instead of running a migration. Marketing lives on its own static SvelteKit site in 12 locales, because the buyers and the users are different people and shouldn't share a bundle.

What shipped was a whole vertical product, not a single app: worker PWA, admin app, marketing site, public API, billing and the hardware, all built together. Workers scan and walk in, managers see the sessions, and payroll carries on running wherever it already ran.