How to save a recipe from any website on iPhone

The short answer

On an iPhone, the fastest route is a Safari extension that spots the recipe while the page is open and saves it in one tap. No link to copy, no app to open first. Everything else on this page is slower, and worth knowing about anyway because you will hit a page where the fast route is not available.

Why recipe pages are like that

It is not laziness and it is not an accident. Two things drive it.

The first is copyright. In the United States, a plain list of ingredients is not protected by copyright. The Copyright Office is explicit about it: a mere listing of ingredients or contents is not eligible. What is protectable is the writing around the recipe, the substantial literary expression, the story about the trip to Tuscany. So the story is the part of the page that legally belongs to whoever wrote it.

The second is money. A page that takes forty seconds to scroll can hold four or five ad slots. A page that shows you six lines and a method can hold one. Nobody is trying to annoy you; they are trying to be paid.

Both of which means the padding is not going away, and the answer is a reader on your side of the screen.

Option one: screenshot it

Everyone starts here. It works until you need to scroll, which is immediately, so you end up with three screenshots and no way to check off what you have already added. The amounts are pictures, so you cannot scale them, cannot search them, and cannot send the list to anyone. And your camera roll fills up with pictures of other people's cooking.

Option two: the share sheet into Notes

Better, because it is text. Worse in practice, because what lands in Notes is the whole page, including the story, the ads that got captured as link text, and the comments. You are back to scrolling, only now in a worse reader.

Option three: paste it into a recipe app

This is what most recipe apps ask you to do. Copy the URL, switch apps, find the import field, paste, wait. It works, and it is four steps too many for something you do while standing at a counter, and every one of those steps is a chance to get distracted and never come back.

Option four: let the page tell the app

Most recipe sites already publish their recipe as machine-readable data, because Google asks them to. It is called schema.org/Recipe and it sits in the page's source as a JSON block: the ingredients, the steps, the times, the yield, all labelled. It is how Google knows to show cook time in the search results.

That means the clean version of the recipe is already on the page you are looking at. A browser extension can read it while you are reading the page, and offer it to you.

This is what Sift does. Turn the extension on once in Safari, and after that a bar appears at the bottom of any page that has a recipe on it. One tap saves the ingredients and the steps with a link back to whoever wrote it. Nothing is republished anywhere. It is a reader for a page you already opened.

What happens on a page with no structured data

Plenty of sites, especially older ones and personal blogs, publish nothing machine-readable at all. A reader that only handles the clean case will tell you the site is unsupported, which is the message that makes people delete these apps.

Sift reads a page four ways and stops at the first one that works:

1. JSON-LD. The schema.org/Recipe block most food sites already publish.
2. Microdata. The older markup, still in the HTML of a lot of sites that never migrated.
3. A readability heuristic. Finds the ingredient list and the step list by shape when nothing is marked up at all.
4. You paste the text in. A screenshot of a cookbook page, a text from your mother, a site that fought all three routes above.

All four land on the same editable card, which is why there is no unsupported-site error state. A clean parse just means you have nothing to fix.

On Android, or on a computer

The Safari extension is iOS. app.siftapp.me opens in any browser and will save and cook a recipe, and browser extensions are the next thing being built.