Building Klau in the Open, Week 12: A Driver App

Week 12 of building Klau in the open.

I built a driver app this week.

That was not in the original plan. Klau was designed to be a layer on top of what haulers already run: make their existing systems better, not replace them. I believed in that and I still do.

But after months of talking to haulers, one thing kept coming up. A lot of them are not happy with what they have. Some of their tools are frustrating. Some are dying on the vine. None of them feel like the obvious answer for the next decade.

I was a driver before I was a developer. And before this week I had already spent most of a decade building driver apps for this industry. I know what works, what doesn't, and what drivers actually need compared with what software vendors think they need. Everything I learned building apps that thousands of drivers use went into this one, built for drivers to use for the rest of their careers.

The app is intentionally thin. No usernames. No passwords. Drivers sign in with their phone number and see their work: today's jobs and a preview of tomorrow. Job states change automatically based on location and movement. The driver confirms completions, photographs dump tickets, and logs lunch. That's it.

Learning the routine

The part I'm most excited about is that Klau learns the driver's routine. If a driver stops at the same deli every day, we expect that and build it into the route plan. If a certain container type takes longer to service, we learn the actual time instead of using a fixed estimate. Self-contained compactors need to be flipped before dumping, and we now have the data to model that correctly.

Routes get better over time. That's the whole idea. Klau is a teammate that learns, not a tool you configure once and forget.

Week 12. Still building.

Originally posted on LinkedIn on March 27, 2026.