The problem
Every travel app is built for plan-time: the desk session where you decide what to do. Almost nothing is built for use-time, the moment you're at a pier or an airport with one hand free, patchy signal and a question that needs an answer in five seconds. Our trip details were spread across email, WhatsApp, Instagram saves, map pins and PDFs, and finding anything was slowest exactly when it mattered.
What I built
A phone-first trip workspace. One screen per day shows what's next and where; tickets and documents open in a show mode for the person across the counter; saved places sit on a map with opening hours and walking distance; cash spends go to Splitwise; and everything keeps working offline.
How I built it
I ran it like a product before writing any code: a product definition with a comparison of existing tools, then a written decision record for each real call (the data model, offline edits, the map, cash). Then I built it with Claude Code in small steps and tested it against our actual itinerary.
How it's used
The four of us used it across nine nights, six stays and eight transit legs. It's kept running between trips.
What's next
Making it work for a second trip without rebuilding anything trip-specific.