Meta Can Crawl Your Website Starting October 30. Decide What It Should See.
Meta's new business terms cover any business that uses Facebook, Instagram or WhatsApp. Here is what changed, what Meta walked back, and the hour of work worth doing first.
By John Martines · October 22, 2026 · 6 min
Meta's revised Commercial Terms take effect October 30 and let it crawl websites tied to any business that uses its products. Meta took the word "train" out after a public outcry, but the crawling stays. Here is what a small business should check this week.
A bakery owner posts a photo of this week's wedding cake on the shop's Facebook page, with a link back to the website. On the website sits the thing she actually sells to other bakers: a paid set of decorating guides, plus a free page of her best recipes that brings in most of her orders.
Starting October 30, the act of running that Facebook page comes with new fine print about that website. Meta's revised Commercial Terms let it send automated software to crawl websites tied to a business that uses Meta's products. There is no separate box to tick. Keep using Facebook, Instagram or WhatsApp for business after that date, and you have accepted them.
Most owners will never read these terms. That's fine. But it's worth an hour to decide what on your site Meta should see, and to set that up before the date arrives.
Meta's Commercial Terms are the rules for any business use of its products. That is wider than it sounds. Coverage of the change says it reaches a company that only manages a Facebook Page, without ever buying an ad. The current version has been in force since March 6, 2024. Meta's own terms page now says the updated version "will go into effect on October 30, 2026."
The new part is a crawling clause. A crawler is a program that visits web pages and copies what it finds, the way a search engine does. Under the revised terms, Meta may crawl websites associated with a business's products or services, check back for updates, and even create "test credentials" to gather information from those sites.
The first version of the clause said Meta could use what it collected to "provide, develop, improve, research, train, and maintain the integrity of" its products. "Train" is the word that matters, because it is how AI models are built. After a day of criticism, Meta rewrote it. The purpose now reads as advertising and business products, and "train" is gone.
How it got here:
- October 15, 2026: Nieman Lab reported the new clause, with legal experts saying it could reach well beyond what most businesses expect.
- October 16, 2026: Meta spokesperson Andy Stone said the update was meant to pull accurate product information from websites, not to scrape news sites, and that nothing had changed about how Meta uses its crawlers.
- October 17, 2026: a revised version of the clause appeared on Meta's legal site, without the word "train."
- October 19, 2026: Search Engine Journal found the UK English preview of Meta's ad terms still carried the original wording, while the US version did not.
- October 30, 2026: the revised terms take effect.
The news coverage focused on publishers, but the terms are the same for a dental office, a contractor or a gift shop. Three things are worth knowing.
The fix didn't close the question. Lawyers quoted in the coverage still read the revised words broadly. One said "develop" is wide enough to include training. Another, a Cornell law professor, called the purposes language "still broad enough to drive a truck through." Nobody knows yet how a court would read it.
Your "no" depends on a voluntary signal. Meta says website owners can still control its crawlers with robots.txt, a small file on your site that tells crawlers which pages to skip. Meta told Nieman Lab the new terms do not give it rights that override those preferences. But robots.txt is a request, not a lock, and a law professor quoted by Search Engine Journal pointed out that it gives a site no enforceable legal right.
Most of your site is meant to be read. Your hours, services and prices are there so people and their AI assistants find you. Being crawled is part of being found. What needs a decision is the smaller set of pages that hold work you sell or that took real effort to make: original photos, detailed how-to guides, recipes, design work, price sheets meant only for existing customers.
1. Make a two-column list of your site. Column one: pages you want everyone to read, people and machines alike (home, services, hours, contact, reviews). Column two: pages that hold work you sell or don't want copied. For most small businesses column two is short or empty. If it's empty, you're mostly done.
2. Put the valuable pages behind a login or off the public site. Anything in column two that truly matters should not sit on a public page. A customer area with a sign-in, a download sent by email after purchase, or a PDF you hand out in person all keep it out of every crawler's reach, not only Meta's. This is the one step that works whatever the terms say.
3. Check your robots.txt with whoever runs your website. Type your web address followed by /robots.txt into a browser to see what yours says today. Meta's developer documentation lists its crawlers by name. Meta-ExternalAgent is the one it describes as crawling for "training foundation AI models." Meta-ExternalAds crawls for advertising and business products. Meta-WebIndexer feeds Meta AI search. Your web person can block or allow each one separately. Go carefully with FacebookExternalHit: it reads a page when someone shares it on Meta's apps, and blocking it may cost your shared links their preview. Meta's page also says two of its crawlers may bypass robots.txt in some cases, which is another reason step 2 matters more than this one.
4. Decide whether the trade is still worth it. For nearly every local business, a Facebook page and Instagram account bring in more customers than these terms put at risk. Leaving in a panic is the wrong move. But if your business sells content, like courses, guides or photography, write down the answer to one question: is what Meta may collect from my site worth more than what my page brings in?
5. Re-read the clause on October 29. Meta has already changed the wording once, and editions in different countries didn't match. Ask whoever handles your ads or social media to open the preview of the terms the day before they take effect and note the date and wording in your files.
- Do we know every website linked from our Facebook, Instagram and WhatsApp business profiles?
- Is anything we sell, or charge for, sitting on a public page?
- Has someone looked at our robots.txt in the last year, and do we know who can change it?
- If we use an agency or a website builder, have we asked how they are handling Meta's crawlers for us?
- Who in the office will notice the next time a platform changes its terms, and where do they write it down?
Every platform that touches your business wants to read your website, for search, for ads and for AI. Meta is simply the one that said so in its contract this month, and the backlash shows how few businesses had thought about it. The businesses that keep their valuable work behind a login, and their public pages clear and accurate, won't have to scramble each time the fine print changes. For the other side of the same shift, how AI assistants describe your business to customers, see our piece [Your Customers May Never Visit Your Website](/articles/mcp-apps-explained).
---
Sources (checked October 20, 2026). Effective dates of the current and revised terms: Meta, "Commercial Terms" (facebook.com/legal/commercial_terms, effective March 6, 2024, with the notice that updates take effect October 30, 2026). Original clause, Meta's October 16 response, the October 17 revision, coverage of Pages without ads, robots.txt statements and expert views from Matt Topic, James Grimmelmann and Benjamin Sobel: "Meta Drops Training From Its Crawl Clause And Keeps The Crawl" (Search Engine Journal, Matt G. Southern, Oct 2026). First report: Nieman Lab, Andrew Deck and Hanaa' Tameez (Oct 15, 2026), as cited by Technology.org, "Meta Drops 'Train' From Website Crawling Terms" (Oct 19, 2026). Andy Stone's statement: Technology.org (Oct 19, 2026) and PPC Land, "Meta grants itself the right to crawl advertisers' websites for AI" (Oct 17, 2026). Meta's crawler names and purposes: Meta for Developers, "Meta Web Crawlers" (undated page, checked October 20, 2026).