You have somewhere north of three hundred recipes saved. Instagram collection, a few hundred screenshots, twenty browser tabs, a note called “recipes” that stops halfway through a curry.

On Tuesday you cooked the same pasta you cook every Tuesday.

Every guide to this problem tells you to pick an app that saves recipes from anywhere. You already save recipes from anywhere. That is the part of your life that is working.

Saving recipes was solved. Cooking from them was not

Every app in this category competes on capture. Save from a website, save from a reel, save from a photo of your mother’s handwriting. It is a real feature and they are all good at it now.

Capture is also the cheap half of the problem. A saved recipe has cost you one tap and bought you nothing yet. The expensive half is the evening you are standing in the kitchen deciding what to make, and no collection of four hundred things has ever made that moment easier.

A recipe collection that only grows is a reading list with pictures of dinner on it.

So the question to ask about a recipe organiser is not how many places it can save from. It is what the app does with a recipe between the moment you save it and the moment you eat it.

A recipe is not saved until its ingredients are separate

This is the whole thing, and it is the reason a camera roll full of dinners never turns into dinner.

A screenshot holds the ingredients as pixels. A note holds them inside a paragraph. In both cases the app cannot tell that the recipe contains four hundred grams of chicken thighs, so it cannot search by what you have in the fridge, cannot scale the quantities when two guests turn up, and above all cannot write your shopping list.

When the ingredients are separate fields, with a quantity, a unit and an aisle, everything downstream becomes free. Put the recipe on Thursday and the shop knows about it. Change your mind on Wednesday and the shop knows about that too.

This is also why scanning is only half a solution. If most of your collection is photographs of pages, adding a recipe from a photo is worth doing properly once, because the text recognition gives you the words and never the aisle.

Organise for six o’clock, not for a bookshelf

Most people file recipes the way a cookbook shop does: Italian, Indian, baking, salads. It feels tidy and it answers a question nobody asks.

At six on a Tuesday you are not browsing by cuisine. You are asking four things, in this order:

1. How long? Under 25 minutes, or not tonight.
2. How much attention? Something that sits in the oven is a different category from something that needs you at the hob for twenty minutes.
3. Will everyone eat it? The single most useful tag in any household with a child in it.
4. Does it do two nights? Which decides whether Wednesday is already sorted.

Tag for those four and your collection starts answering questions instead of storing answers. Cuisine still earns a place as a second layer, for the Saturday when you feel like browsing. It is a bad primary index.

Getting the backlog in, without losing a weekend

Nobody is going to transcribe three hundred screenshots, and you should not try. The backlog is mostly aspiration anyway.

Do it in the other direction.

• Start with what you already cook. Write down the fifteen or twenty meals your house eats without looking anything up. Those are the recipes that matter and most of them are not saved anywhere.
• Then add one at a time, at the moment of cooking. The evening you actually make the thing from the screenshot is the evening it earns a proper entry.
• Leave the rest where it is. The Instagram collection can stay an Instagram collection. It is a wish list, and wish lists are allowed to be messy.

Six weeks of that and you have a real recipe box, built out of evidence rather than intention. It also means the tedious part is spread across six weeks in one-recipe pieces, which is the only reason it gets done at all.

Count cooks, not saves

Here is the number every recipe app shows you: how many recipes you have saved. Here is the number that would actually change your Tuesdays: how many times you have cooked each one.

Almost nothing tracks the second, because it needs a button somebody remembers to press, and nobody presses it. The way around that is to derive it. If you put a meal on a day and that day arrives, you cooked it. No extra tap, no honesty system.

Once you have that number, sort the box by it. The top of the list is your household’s real menu, and it is usually shorter and more boring than anyone expects. That is useful information rather than an indictment.

What the number is for

Recipes cooked nine times are the rotation, and they should be the easiest ones to reach. Recipes cooked once were a phase. Recipes saved eighteen months ago and never cooked are answering a question about who you would like to be, which is fine, and they should not be in the way on a Tuesday.

The number is also the only honest way for an app to suggest anything. Chirpy’s Kitchen uses it for one nudge and nothing more: you have made this nine times, while that one has not come up in a while. No taste model, no guessing what you feel like. Just the two ends of a list you built by cooking.

Twenty recipes is a full collection

Four cooking nights a week against twenty recipes is a five-week rotation. Nobody at your table can detect a five-week rotation. They can detect the same three meals, which is what happens when a collection is too big to choose from and you fall back on memory.

This is the same shape as planning four cooking nights instead of seven: the smaller number is what makes the system survive a bad week. Keep six that take under 25 minutes, six standards nobody argues about, four that make good leftovers, and four for weekends.

Everything else lives in a second list called “to try”. One graduates each week or it does not. No pressure either way.

Paprika, a notes file, or the app your house already opens

Three honest options, and they fail in different places, which is the part worth knowing before you install anything.

If all you want is a recipe box with a good web clipper, Paprika Recipe Manager 3 is $4.99 on the US App Store as a one-off, checked on 2 September 2026, and Paprika’s own site notes that each platform version is sold separately. For a single cook on a single device that is very hard to argue with, and we would rather say so than pretend otherwise. For a house with mixed phones it is where the cost starts, which is what our guide to picking a Paprika recipe manager alternative works through, including what changes when version 4 arrives.

The case for keeping recipes inside the household app is about the other people in the house. A recipe box that lives beside the calendar and the shopping list is one your partner already has open for other reasons, which matters more than any feature, for the same reason most shared family calendars go unread.

The last hop: recipe to shop

A recipe collection earns its keep at the supermarket. Four recipes picked on Sunday should produce one list, deduplicated, in aisle order, visible to whoever gets there first.

That hop is where most setups leak, and the leaks are boring and specific: two recipes both want garlic and it goes on twice, the list is sorted by recipe so you cross the shop four times, one person adds to the list in a different app. It is the same failure pattern as the four places a shared shopping list quietly leaks, arriving one step earlier in the chain.

The fix is structural rather than diligent. If the week’s plan can build the list in one action, nobody has to be careful, and the garlic gets merged whether or not anyone was paying attention.

Common questions