Your recipes, credited and celebrated.

Why this matters more than it used to

A recipe travels; your name usually doesn’t. It gets screenshotted, retyped, and reposted until the version ranking on Google belongs to a site that had nothing to do with it. We watched one recipe turn up on nine near-identical content farms before the creator’s own post was a year old.

That mattered when people searched. It matters more now that they ask. When someone asks an assistant “who makes good weeknight pasta?”, it answers from whatever it can read and resolve — and a handle buried in a video caption is close to invisible to a machine.

What we actually do about it

• Every saved recipe becomes a readable record. Each public recipe gets its own page carrying schema.org Recipe markup, and the author field holds your name and a link — not a bare string a machine can only guess at. The recipe also cites your original post.
• You get a page of your own that a crawler can actually read. Your collection lives at thepantrybutler.com/creators/your-name as a schema.org ProfilePage describing you as a Person, with every recipe cooks have saved from you listed on it. It is server-rendered, so it does not need JavaScript to be seen.
• You become an entity, not a string. The Instagram, TikTok, YouTube and website fields above go into sameAs — the field that tells a machine those accounts are all the same person. That is what turns “@yourhandle” into something an answer engine can match across platforms and cite with confidence.
• All of it is in the sitemap. 1,782 URLs today, 290 of them creator pages. Nothing is hidden behind a login or a tab that only renders in a browser.
• Assistants can also read us directly. We run a server on the open Model Context Protocol, listed in its public registry, so an assistant can query recipes with their credit attached rather than scraping a page and hoping.
• It compounds. Every cook who saves one of your recipes adds another attributed record pointing back at you.

What we have actually measured

We count every AI crawler that fetches a page here, because a claim about being readable is worth nothing if nobody is reading. This is the whole log, not a highlight:

Real cooks, not content farms

The reason a recipe loses its author is that reposting it is profitable. An account lifts a dish, strips the credit, pads it with ad copy, and outranks the person who cooked it. The nine-farm recipe above is not a story we read somewhere — the account behind it sat at the top of our own directory, by saves, until we looked properly.

So we look. Before an account appears as a creator here, it has to clear three things: a named human behind it, original work rather than reposts, and a catalogue rather than one lifted viral hit. Follower count is not on that list and never will be — the worst account we found had thousands of real followers, and the best had ten people who cooked its food.

When an account fails, we hide it from Discover and write down why, with the evidence and the date. Three are hidden so far. What we do not do is delete the recipes: someone saved those to cook them, and that person’s library is not the thing at fault. The credit stops being promoted; the cook keeps their dinner.

That is the whole product argument in one sentence: credit should survive the trip. Everything above is just the plumbing that makes it true for machines as well as people.

What we will not tell you: that this makes an AI recommend you. We do test it — we have run 5 checks asking assistants questions our recipes should answer, and we have been named in 0 of them. Being crawled is not being cited, and we would rather show you that number than hide it. What we can say is that your work here is machine-readable, correctly attributed, linked back to you, and not sitting next to a farmed copy of itself — which is more than it gets almost anywhere else. Figures on this page were pulled from production on 21 September 2026.