How we crawl

PlatebookBot is the crawler for Platebook — a meal planner for people whose food needs are specific. Food allergies, celiac, Crohn's, gastroparesis, dysphagia, PKU, and a few dozen more, where the ingredient list is what decides whether a dish is an option.

If PlatebookBot is in your logs, this page is the whole story: what it fetches, what Platebook publishes afterward, and how to switch it off.

Last updated 4 August 2026. Anything on this page that changes gets a new date here.

Identifying us

Every request carries this User-Agent:

PlatebookBot/1.0 (+https://platebook.co/bot; recipes@platebook.co)

That's the only signal we currently offer, and a User-Agent is trivial to forge — there's no published IP range to check us against yet. If something calling itself PlatebookBot behaves in a way this page doesn't describe, it isn't us. Send the log lines to recipes@platebook.co and we'll tell you either way.

What we fetch

  • Recipe pages, to analyze them. Platebook checks a recipe against allergen presence, condition compatibility, texture, and effort — so someone with a sesame allergy, or in a Crohn's flare, can find out whether the dish works for them before they start cooking.
  • A copy of the HTML, held while that analysis runs. Analysis happens in stages, and keeping the page we already fetched means a later stage doesn't send us back to your server for the same bytes.
  • Nothing in bulk without a person choosing to. Discovery collects candidate URLs; crawling them takes an explicit approval from someone at Platebook. There's also a per-domain switch that halts work already in flight.
  • Slowly. Crawl-delay is honored, capped at 30 seconds — long enough to respect a busy server, short enough that a stray Crawl-delay: 86400 doesn't silently park a queue for a day.

What we publish

  • Our analysis, your ingredients, your name, and a link to you. A Platebook recipe page shows the compatibility verdicts we derived, the ingredient list, and a prominent link to your site. Your instructions, your photos, and your writing stay on your page — and that's where every cook gets sent.
  • Your photo, from your server. Where a thumbnail appears it's your og:image URL, loaded from you and linked back to you. No copies at full size, ever.

What we honor

  • robots.txt, with Allow/Disallow precedence per RFC 9309. Disallow us and we don't crawl you.
  • Crawl-delay, capped as above.
  • 403s and bot challenges. A site that refuses an identified crawler has answered, and we don't go looking for another way in.
  • Content Signals — not yet. Content-Signal parsing is being built and isn't live. We'd rather say so than let you assume a control is working. Until it ships, robots.txt is the one that is.

What we never do

  • We never train or fine-tune AI models on your content. Not ours, not anyone's.
  • We never republish your instructions. No Platebook page shows your steps as something to cook from — not copied, not summarized, not rewritten in Platebook's voice.
  • We never present your recipe as ours. Your name and your link are on every listing that came from you.
  • We never evade a block. No user-agent spoofing, no challenge circumvention, no proxy rotation.

Turning us off

Two ways. Both hold until you say otherwise.

  1. Add this to robots.txt:

    User-agent: PlatebookBot
Disallow: /

  2. Or email recipes@platebook.co. We'll stop crawling and pull your existing listings within 7 days, then write back to confirm it's done.

You don't have to give a reason, and we won't ask for one.

Working with us

The opposite direction is open too: your recipes on Platebook in full, instructions and all, in front of readers searching by the exact constraints your food happens to meet. That happens only by agreement, never by default. Email recipes@platebook.co with "partner" in the subject line.

Anything here that doesn't match what you're seeing: recipes@platebook.co.