Projects / Retail & discovery · Useful tools
Exit Pass
What if grocery checkout were finished before you got to the front of the store?
A working prototype that runs entirely on simulated store systems. It is not connected to any store.
- State
- Simulation
- Last change
- Oct 2, 2026 · latest commit
- Page last reviewed
- Oct 29, 2026
- Built with
- AI coding assistants. Commit trailers show Claude as a co-author.
The problem or experience
Scan-and-go apps still end at a register where someone audits the cart. Exit Pass explores targeted verification: you pay as you shop, and the exit check only looks at the one to three things that matter, so most trips skip it.
Who it's for
Shoppers who hate lines, and store teams and builders who think about checkout, shrink and staffing.
My contribution
It is my project. The shopper phone app, the register lane, the pass decision engine and the store radar are built from scratch against a simulated point-of-sale.
What works today
- A shopper flow with scanning, a running tax-included total and a live pass QR.
- A register lane that recalls a paid basket and handles the check.
- A store view that shows carts arriving, with simulated shoppers.
Limitations
- Point of sale, payments, cart sensors and store traffic are all simulated.
- The README's security notes say what a production version would still need.
- I haven't found a hosted demo. You run it locally with npm.
Evidence
- README, security notes and source · checked Oct 29, 2026
Updates
Exit Pass: explaining spot checks to shoppers
Exit Pass now tells shoppers why a random spot check happened, and gives register lanes an actionable hint when their sync is out of date.
How you can get involved
Work in retail or build checkout software?
- Tell me where the pass logic would fall apart in a real store.
- If you want to pilot something like this, I'd like to talk.