The Grocery List Prompt That Ends Food Waste

One message turns a week of dinners into a shopping list grouped by supermarket aisle, with anything you already own moved to a separate CHECK FIRST section and every single-use ingredient starred. It takes about four minutes and it’s where the money in meal planning actually is.

The prompt is below, then the four lines inside it that do the work — because three of them exist to stop you buying food you’ll throw away.

This is the shopping half of how we plan meals with AI. You need a plan before you need a list, and that page has the whole loop.

The prompt

Send this straight after the model has given you a plan you’re happy with. It works on whatever’s already in the conversation, so there’s nothing to re-paste.

The four lines, and what each one prevents

Group it by aisle — so you walk the shop once

A list in the order the meals were planned sends you back and forth across the shop, and every extra lap is another pass down an aisle you didn’t need. Grouped by aisle you walk it once and leave.

Name the sections rather than saying “group by aisle” and hoping. Produce, dairy, meat and fish, dry goods, freezer, other — six is enough, and it roughly matches how most shops are laid out. If yours is different, say so once and it’ll use yours.

Quantities the way a shop sells them — so you can buy it

Left alone it’ll give you recipe quantities: two cups of this, 150 grams of that. Nothing is sold in cups. You end up standing in front of the tinned tomatoes doing arithmetic, which is exactly the kind of small decision this whole system exists to remove.

A bunch, a tin, a 500g pack, one loaf. Things you can put in a trolley without converting anything.

CHECK FIRST — so you stop buying your third jar of paprika

Anything you mentioned already having goes into its own section at the bottom rather than being silently dropped from the list.

That distinction matters more than it looks. Dropped items are invisible, so you can’t verify them — and half the time the thing you said you had is a quarter of a bag, not a bag. A short CHECK FIRST section is a thirty-second look in your own cupboard before you leave the house, which is the cheapest thirty seconds in the whole process.

Star the single-use ingredients — the actual waste mechanism

This is the one to keep if you keep only one.

Half a bunch of dill for a single Tuesday. A jar of something for one tablespoon. A vegetable nobody has cooked with before, bought for one meal on a Thursday that’s going to become a takeaway anyway. Individually none of them is much. Together they’re most of the gap between what a week should cost and what it does.

The star doesn’t decide anything for you. It turns an invisible cost into a short list you look at before you leave. Sometimes the answer is buy it. Sometimes it’s ask for a swap. Either way you chose.

What the list looked like

Five dinners in, one list out, in the six aisle sections asked for — produce, dairy, meat and fish, dry goods, freezer, other. Dairy came back as None rather than being dropped, which is the right answer.

Every quantity is something a shop sells: 1 kg potatoes, 600g chicken, 2 tins chickpeas, 12 eggs, 1 bulb garlic, 1 small bag frozen peas. Not one cup. Nine items came back starred, with a legend explaining what the star meant, and the CHECK FIRST section held the spinach, mince and pasta that had been named as already in the house.

It also put lentils in CHECK FIRST, and nobody had said they owned lentils. Thursday’s dal needs them and they are not on the buy list — so if you did not have them, Thursday quietly fails. Read the CHECK FIRST section as a list of claims to verify rather than a list of things you have. That is what it is for, and this is what happens when you skip it.

Shop to the list, not to the shop

The plan is rarely what fails. What fails is standing in the shop adding things to the trolley.

Take the list on your phone, buy what’s on it, leave. Everything clever has already happened — the aisle order, the quantities, the starred decisions you made at home where there was no one behind you. The shop is just execution.

This is also why the list is worth generating even in weeks when the plan half falls apart. A list you shopped to is a week of food in the house. A plan you didn’t shop for is a piece of paper.

When something isn’t there

Half the shopping-list advice on the internet stops at the moment the list is printed, which is roughly the moment it stops being accurate. Things are out of stock. Prices aren’t what anyone assumed. The pack size is 400g and the list said 500g.

Ask for the swaps before you go, not in the aisle

One extra message while you still have the conversation open:

For each item on this list, give me one substitute that would work in the same meal. One line each, no explanation.

You get a second column you can glance at in the shop. It costs about thirty seconds and it’s the difference between a swap you chose and a swap you improvised while somebody waited behind you.

Ignore the prices

If you asked for costs, you’ll have got them, and they’re a guess wearing the clothes of a figure. It doesn’t know what your shop charges, which week the offers changed, or what a pack size is where you live.

What it’s actually useful for is the shape of the spend rather than the number. Which three items are doing the least work for their price? is a question it answers well, because that’s a judgement about how many meals an ingredient touches — which it can see — rather than a number it can’t.

Pack sizes are its blind spot

It’ll ask for 300g of something sold in 400g packs, every time, in every country. That’s not a fixable flaw and it isn’t worth a prompt line — just buy the pack. The leftover is what Friday’s dinner is made of, and there’s a section about that below.

What’s left on Friday

By the end of the week there’s usually an odd collection of things. This is the job these tools are genuinely best at, because there’s no guessing involved — you’re telling it exactly what exists. The fridge clean-out prompt is that job with the four destinations a leftover usually has.

Type what’s there in plain words. I have half a cabbage, three eggs, and some rice. Dinner for two, nothing else bought. That’s the whole prompt. No structure needed, because the constraint is doing all the work.

Do that once a week and the starred-ingredient problem mostly solves itself from the other end.

Cooking down the pantry, on purpose

Everything above works on a plan that already exists. The stronger version runs the other way round: build the week from what’s in the house first, then buy the gap. That is the pantry audit, and it caps the shop at about five items.

List what’s actually in your cupboards and freezer — not a full audit, just the things that need using — and add one line to the planning prompt: build as much of the week as possible from this before adding anything.

The list that comes out the other end is shorter, cheaper, and has almost nothing starred on it, because the meals were designed around what you owned rather than the other way round. It takes two extra minutes and it’s the single biggest change you can make to what a week costs.