Meta's New Business Terms Take Effect October 30 and Let It Crawl Your Website. After Backlash It Dropped One Word, and the Crawl Right Stayed.
Meta's revised commercial terms take effect on October 30, 2026, and they add a clause that lets Meta, or anyone acting for it, run crawlers against the domains associated with your products or services. If you buy ads, manage a Page, or use Ads Manager or Business Suite, you count as a business customer. Meta has reportedly removed the word "train" from the permitted uses after backlash, but the crawl permission is still there. You have until the end of next week to decide what your site should say about that.
What the clause says
According to The Register, the new language authorizes Meta to use "web scrapers, bots, robots, spiders, crawlers, and user agents," and to "create test credentials," to access and crawl domains, web pages, or URLs associated with a customer's products or services. It also covers periodic pulls of updates. Meta may use what it collects to "provide, develop, improve" its products, services, and features. SC Media's summary of the original wording lists a longer set of purposes that included "train."
Meta defines business customer broadly: advertisers, sellers, app developers, Page and business Group managers, and people using Meta's business products. The Register reported that Meta did not respond to a request for comment on the original terms.
What changed after the backlash
Quartz, citing Nieman Lab, reported that Meta removed the word "train" from the permitted uses of scraped content around October 16. A separate writeup says the change applied to the US English version and that the UK English version still listed "train" when checked on October 19. I could not confirm that second detail from a primary source, so treat it as unverified. What every account agrees on is that the crawl permission itself stayed, and legal experts quoted by Quartz say the revised terms could still let Meta scrape sites tied to business accounts.
The "develop, improve" language is where I would focus. Removing one verb from a list does not change what a crawler can fetch, and "develop" and "improve" are wide enough to cover most of what a model team wants to do with web text.
robots.txt helps, but the contract doesn't promise it
Reports say Meta intends to let businesses signal crawling preferences through robots.txt. Eric Goldman, a Santa Clara University law professor, noted in The Register's coverage that the terms do not say Meta "will respect those robots.txt instructions." The News/Media Alliance says there is no way to opt out without ceasing to use Meta's platforms.
So robots.txt is a request, backed by a company's public statements rather than by your contract. One writeup of Meta's crawler documentation lists meta-externalagent as the crawler used to train foundation models or index content, and says Meta states it respects robots.txt. The same writeup lists facebookexternalhit as the link-preview fetcher that can bypass robots.txt for security checks.
What I'd actually do before October 30
I would make four changes, in about twenty minutes.
First, add a block for Meta's AI crawler to your robots.txt:
User-agent: meta-externalagent Disallow: /
Second, do not block facebookexternalhit. That fetcher builds the link previews for your posts and ads, and blocking it breaks the thumbnails people click on.
Third, audit what is public on the domain you have attached to your ad account or Page. If a pricing page, a template library, or a paid guide's preview is readable without a login, assume it is crawlable. If you rely on a free signup wall to protect premium content, note that the clause explicitly mentions creating test credentials, so a free account is not much of a barrier.
Fourth, decide whether your Meta spend justifies the exposure. If Meta ads bring in most of your customers, you will accept the terms and block what robots.txt lets you block. If they bring in 5% of them, the cleaner move is to stop running ads against the domain that holds your best content.
Where I could be wrong
An honest reading of the revision is that crawling a public site for ad and business products is closer to what Google and others already do than to a land grab. If your site is mostly marketing pages you already want indexed, the practical exposure may be close to zero. My concern is narrower: you are agreeing to a contract term that is broader than the technical controls Meta says it will honor, and the gap between the two is yours to carry.
Author
Lukas
@lukcombinatorSources
- Meta grants itself right to scrape content from business customers for model training
- Meta removed AI training wording from business terms after backlash
- Meta's new terms allow AI training on business customer content
- Meta Commercial Terms 30 October 2026: Meta Can Crawl Your Business Website and Train AI on It, and How to Block It