Help

About twenty minutes for a shop with thirty or so items, and most of that is deciding what you actually want to count rather than typing.

Start with the ten things that cost you the most or run out the most often — milk, beans, cups, pastries. You are not trying to inventory the whole shop on day one, and a list of ten items that gets counted every morning is worth more than a list of eighty that gets counted once.

If you already have a stock list anywhere — a spreadsheet, a photo of a sheet on the wall, a supplier order form — upload it and it will read it and propose the items for you to check before anything is created.

No. Everything staff-facing runs on the phone already in their pocket, using the normal camera. There is no app to install and no hardware to buy.

The only thing you need is a printer for the QR labels, and an ordinary office printer on plain paper is fine. Sheet labels and thermal label printers are supported if you happen to have one.

The spreadsheet is usually fine at recording. What it can't do is be in two places at once — the person who notices the milk is low is behind the bar, and the spreadsheet is on a laptop in the office.

The practical difference is that counting stops being a job somebody sits down to do. It happens where the stock is, in four seconds, by whoever is standing there.

You can paste the spreadsheet in to start from what you already have.

No, and deliberately so. A barista scans the code, enters the shop PIN once on that phone, types their first name once per shift, and from then on it is one tap.

Nobody has to be invited, remember a password, or be removed when they leave. The people who need accounts are the ones who look at the numbers, which is usually you.

A four-digit code that proves a phone belongs to someone who works in your shop. It is entered once per phone, not once per scan.

You set it when you create the shop, and you can change it in Settings. Changing it does not log anybody out — phones already trusted stay trusted, which means you can change it after somebody leaves without disrupting the morning.

Mark them inactive on the Team tab. Any availability link they were sent stops working immediately, their name stops being suggested on the scan screen, and every record they made is kept exactly as it was.

If they had their own owner-level login, deactivating it takes effect at once — they cannot get back in using the shop password either.

Change the shop PIN too if you want to be thorough. It doesn't sign anybody else out: phones that have already been trusted stay trusted, so you can change it after somebody leaves without disrupting the morning.

The count is held on the phone and sent when the signal comes back. Staff see that it is waiting rather than a failure, and nothing is silently lost.

This covers the ordinary case of a thick wall or a dead spot by the walk-in. It is not a full offline mode: a phone that never reconnects will still be holding the count.

Add items to the print queue from the Items page, then choose a format. Plain paper works; so do Avery sheet labels and Dymo or Brother label printers if you have one.

Print at 100%. A printer set to 'fit to page' shrinks the code, and a shrunk code on a small label is the one that will not scan on a dark morning.

Print a wall card and stick it where the job happens. One by the door for opening and closing, one on the grinder, one on each fridge. Anyone points a phone camera at it and the right list opens.

When you print, you choose what the card is for. A card stuck on a specific thing always opens that thing's checklist. A card by the door opens whichever list is due right now — the opening checks at 7am, the closing ones at 5pm — so one card covers the whole day.

Laminate them, or use a plastic sleeve. They live next to a sink and the uncovered ones come down inside a month.

Every printed label carries a short backup code underneath the square. Staff can type that instead and carry on.

Reprint the label when you get a minute. The code on the new one is the same code, so nothing about the item's history changes.

On the shelf, for almost everything. A label on a tub leaves with the tub, and the first empty one that goes in the bin takes the code with it.

The exception is something that always lives in one container and gets refilled — a bean hopper, a syrup bottle. Put it on the container there.

Whatever you will actually keep up. Most shops that get value from this count the fast-moving things daily at close — milk, beans, cups — and everything else weekly.

Consistency matters more than frequency. Counting the same things at the same time each day is what makes a change mean something; counting everything once a fortnight tells you almost nothing about why.

Yes. Open the item's history and void the count. It stays on the record marked as voided with who voided it, and stops affecting the figures.

Nothing is deleted outright, on purpose. A stock record you can quietly edit is not evidence of anything, and the same principle is what makes the food-safety records worth showing an inspector.

Set that item to be estimated rather than counted. Instead of typing a number, staff tap one of four buttons: full, three-quarters, half, quarter or empty.

It is less precise and much more likely to actually happen, which on balance makes it the more accurate of the two.

You set a par level on the items you care about — the amount you want to have on hand. Anything below its par appears on the ordering list, grouped by the supplier you usually buy it from, with an estimate of what the order will cost based on what you last paid.

It is a suggestion, not an automatic order. Nothing is ever ordered on your behalf.

Yes. Put the address of the page you normally log into on the supplier, and it becomes a button next to everything you buy from them on the ordering list.

Individual items can carry their own link too, for the handful of things where ordering the wrong size has cost you before.

No. Photograph the invoice and it reads the lines and the prices, matches them to your items, and shows you what it found before anything is saved.

It is best-effort and you should look at it. A creased thermal receipt in bad light is genuinely hard to read, and the review screen exists because of that rather than as a formality.

Yes, and it looks for the slow version as well as the obvious one. Suppliers rarely raise a price by fifteen percent in one delivery; they raise it by two percent, four times a year.

It compares the start of a period with the end rather than one delivery with the last, so a price that has crept up over a quarter shows up as a price that has crept up over a quarter. It appears in the Monday email, with what the difference costs you at your current usage.

It is what you had, plus what was delivered, minus what you have now — set against what your recipes say should have gone out, minus what was logged as waste. Whatever is left over is unaccounted for.

That gap is usually over-pouring, an untrained hand on the grinder, drinks going out unrung, or breakage nobody logged. It is the difference between knowing your margin on paper and knowing what you actually made.

It needs an opening count to mean anything, and it says so rather than guessing when there isn't one.

Three. What your ingredients cost as a share of what you take, what your staff cost as a share of it, and how much each sale is worth. The first two added together is the figure most operators watch.

All three are on the Numbers page against the ranges typical for independent shops, so a figure that is drifting is visible while there is still time to do something about it.

Where a figure cannot be worked out honestly — no menu prices set, no wages entered — it says so rather than showing a confident number built on a gap.

Only someone signed in with an owner-level login. Wages are never shown on a staff-facing screen, never included in a scan or a checklist, and never sent to anybody's phone.

You can leave a wage blank. Those hours are then counted as unpriced rather than free — the labour figure says it is incomplete instead of quietly reading lower than reality.

That is what it is for. Every temperature reading and every completed checklist carries what was checked, when, by whom, and what was done about anything that failed. Export ninety days as one file whenever you need it.

Two rules make it worth showing. Nothing can be edited after the fact, and a failed reading cannot be saved without saying what was done about it — which is the part an inspector actually asks about.

Yes — rename it, add or remove steps, reorder them, change how often it is expected, or archive it entirely.

Records already signed off keep the steps they were signed off against. Editing today's checklist cannot rewrite what somebody ticked last month, which is exactly why the records are worth anything.

Change it. Each unit has its own range and its own number of checks a day, and both can be edited at any time.

Leave one side empty for no limit — a freezer has a maximum and no minimum. Readings already logged keep the result they were given at the time.

Square connects directly. You sign in to your own Square account, match your menu to your recipes once, and from then on completed orders draw stock down on their own. It only reads — no write access, no payments, no customer details.

Clover and Toast do not have a live connection yet. Those shops export a sales file from their own dashboard and upload it, which takes a minute and brings in the whole history. You get the numbers; you do not get automatic stock depletion.

No till at all is fine too. Stock, waste, ordering and records all work without one.

Read completed orders. That is the whole of it. It cannot take payments, issue refunds, change your menu, or see customer names or card details.

You can disconnect it from your own Square dashboard at any time without asking anybody.

Send them a link. They open it, tick the days and hours they are free, and that is it — no login, no account, no app.

Their availability then shows on the rota you are building, so a shift that falls outside it is flagged while you are placing it rather than the morning it goes wrong.

It warns, across every location you run, and it does not block. Real scheduling sometimes needs an exception and software that refuses one just gets worked around.

It also flags shifts outside somebody's stated availability, and hours that look like too many in a row.

Yes. Connect a Google Calendar and every shift you schedule appears on it, so staff see their hours without logging into anything.

Changes made in the app push to the calendar. Changes made directly in Google are picked up on a daily sync rather than instantly.

Counter is free permanently — scan-to-count stock, par levels and the reorder list, waste priced at what you paid, printable labels, and a full export whenever you want one.

Shop is $39 per location per month. Everything one cafe runs on. Group is $89 per location per month. For more than one address.

Yearly billing is ten months for twelve.

21 days, and no card is asked for to start it.

If the trial ends and you have not added one, the account drops to Counter and keeps every record you made. Nothing is deleted, nothing is locked behind a payment screen, and nobody is charged for a trial they forgot to cancel.

In Billing, yourself, without emailing anybody. You keep access until the end of the period you have paid for, and then the account drops to the free plan with every record intact.

There is no retention flow and nothing to click through. If it is not worth the money it is not worth the money.

Suppliers, recipes and staff are shared across the whole business; stock, counts and records belong to each shop. You switch between them from the header, and Group adds one view across all of them.

Somebody can be assigned to more than one site, and the rota checks for clashes across all of them rather than one at a time.

All of it, as one file, any time, without asking. Items, counts, sales, waste, deliveries, recipes, temperature readings and checklist records.

This is deliberate and it is the answer to the question underneath the question: you are not locked in, and nothing here works by making it difficult to leave.

Yourself, in Settings, without contacting anybody. It asks for your password, your shop's name typed out, and a one-time code emailed to you.

Three steps because it cannot be undone and there is no backup you can reach afterwards. Export first if you want a copy.

You, and anybody you give an owner-level login to. Nothing about your shop is shown to another customer, and there is no shared or benchmarked view built from your figures without your say-so.

The benchmarks on the Numbers page are published industry ranges for independent cafes, not other customers' data.

Each shop's data is isolated from every other shop's, and that isolation is tested automatically on every change — a suite that signs in as two separate businesses and tries every way one could reach the other's records. All of it has to fail.

Passwords are hashed, sessions are signed, third-party tokens are encrypted at rest, and there is an off-site encrypted backup.

It is also early software run by one person, and that is worth knowing alongside the rest of it.