A coffee cart is the leanest business in food service. One person, one machine, a folding table's worth of pastries, and a line that forms all at once. Nearly every "best coffee shop POS" guide you'll find was written for a different business than yours — a shop with a counter, a register, and a second employee standing at it. Square, Clover, and Toast are all good at that job. But they all assume the same thing: that a staff member rings up every sale.
On a one-person cart, that assumption is the problem.
The bottleneck isn't the coffee. It's the second transaction.
Watch a solo cart during a rush and you'll see it: the drink is nearly ready, and three people are now standing there holding a scone and a bottled cold brew, waiting for you to stop, wipe your hands, pick up the reader, tap through a menu, and take their money — for items you didn't even have to make. The espresso is not the constraint. The payment is. Every retail item and every simple pre-made drink has to queue behind your two hands.
The fix isn't a faster reader or a bigger menu. It's removing yourself from the transactions that don't need you.
What "self-checkout" means at a cart
A self-checkout POS turns the customer into the cashier for the things they can handle themselves. It runs on the phone or tablet you already own, mounted at the edge of the cart. Here's the whole loop:
To be clear about what this is not: the camera does not silently watch your table and charge people as they pick things up. The customer chooses the item and taps to pay — it's a checkout they run, not a surveillance system. That distinction matters, because the "just walk out" fantasy needs ceiling cameras and a sensor-lined shelf. A cart needs neither.
Pastries don't have barcodes. That's fine.
The awkward part of any cart catalog is the baked goods. A croissant has no UPC. A muffin from the wholesaler down the street has no UPC. This is exactly where the generic grocery-style scanner falls down and where people give up and just keep a cash tin.
TallyTill handles the no-barcode item three ways, and the customer picks whichever is fastest:
- Tap the tile. Your daily pastries live as big buttons on the screen. One tap adds the almond croissant to the cart. For a cart with a dozen SKUs, this is usually the quickest path.
- Snap a photo. For the item they can't find, the customer takes a picture and server-side AI matches it against your catalog. Useful when the menu rotates and you don't want to reprint tiles every morning.
- Scan the barcode. For the packaged things — bottled water, canned espresso, a bag of whole beans — the built-in scanner reads the barcode on-device, instantly.
You build the catalog once, from your phone, and adjust it in seconds when the bakery sends something different.
Most of your line isn't carrying cash
If your cart is cash-or-Venmo, look at what your customers actually reach for. Across all U.S. consumer payments, cash has fallen to a small minority of transactions — cards and phones are how people pay now.
Roughly two out of three payments are cards, and mobile-phone payments keep climbing every year. A cart that can only take cash isn't just inconvenient — it's turning away the majority of the way people want to pay, one impulse pastry at a time. A self-checkout that leads with tap-to-pay and a QR code meets the line where it already is.
When the signal drops in the park
Carts work where the Wi-Fi doesn't: a park, a festival lot, the far corner of a market. TallyTill is built for that. The catalog, the barcode scanner, and the cart all run on the device itself, so a weak signal never freezes the checkout. QR payment clears through the customer's own phone and bank rather than your connection, so it goes through even when your bars don't. For card taps, the till can hold and settle the charge once you're back on a solid connection — so a dead zone costs you a delay, not the sale.
Where a self-serve lane helps — and where it doesn't
Honesty matters here, because self-checkout is not a universal replacement for you standing at the cart.
- It shines for grab-and-go retail — pastries, bottled drinks, bags of beans, merch — and for simple pre-set drinks like a drip coffee or a canned cold brew. These are the sales clogging your line for no reason.
- It also works as order-ahead-at-the-cart: a customer rings up and pays for a latte themselves, and the ticket appears for you to make. They've already paid, so you hand it over and move on.
- It won't pull your shots or steam your milk. A made-to-order drink still needs a barista. The point isn't to remove you from craft — it's to remove you from cashiering, so more of your attention goes to the cup.
The right mental model is a second lane. You keep making drinks; the tablet handles everyone who's only there for a scone and a tap.
How it's different from Square, Clover, and Toast
Those systems are genuinely good, and if you want a traditional attended register they're a fair choice. The difference is structural, not a feature checklist:
- A standard cafe POS is operator-run. Every sale flows through the person holding the device. Add a second lane and you've added a second employee.
- A self-checkout POS is customer-run. The second lane is the customer's own phone and your mounted tablet. You add capacity without adding payroll — which, for a one-person cart, is the only kind of capacity you can afford.
You're not choosing a better register. You're choosing whether the register needs you at all.
Setting it up on a phone you already own
There's no dedicated kiosk to buy. The practical version of getting started looks like this:
- Install on a spare phone or tablet and mount it at the edge of the cart, facing out.
- Build a small catalog — your regular pastries as tiles, your packaged items by barcode. Ten minutes for a typical cart.
- Turn on the payment methods you want to accept: contactless card, mobile wallets, QR.
- Set a tip prompt if you use one, and you're taking self-serve sales.
Because it runs on hardware you already have, the whole thing is reversible. If a market morning is slow and personal, ignore the tablet and ring people up yourself. When the rush hits, it's a second lane waiting.
Start with the free plan
If you run a cart, the cheapest way to find out whether a self-serve lane helps is to try it on a real Saturday. TallyTill's Stand plan is free — there's no monthly seat cost and no card required to create an account — so you can build your catalog and take sales without committing to anything. See the current plans and per-sale details before you scale up to more carts.
Create a free account and set up your cart's catalog before your next market.