On October 30, 2026, Meta’s Commercial Terms change for every business that runs a Facebook Page, an Instagram account, or a WhatsApp Business profile, whether or not that business has ever paid for an ad. The new clause authorizes Meta to deploy automated software, web scrapers, bots, crawlers, and even create test login credentials, to reach the websites tied to a company’s products or services. Before March 2024, nothing like this existed in the Commercial Terms at all; the same language lived only in the separate agreement signed by advertisers.

What the new clause actually authorizes
The text itself, read directly on Meta’s legal page this week, sits under the heading “Licenses and Rights You Provide to Meta.” It authorizes Meta to use automated software, including web scrapers, bots, spiders, crawlers and user agents, and to create test credentials to access domains, webpages or URLs connected to a business’s products or services, then to index, cache, analyze, collect, retrieve or crawl that content and pull periodic updates. None of this hinges on an ad budget. A company that only posts to a Facebook Page, runs a WhatsApp Business catalog, or manages a product listing inside Meta Business Suite now falls under the same clause as a paying advertiser. Acceptance is passive: continuing to use any of those products after October 30 counts as agreeing, with no checkbox and no partial opt-out documented anywhere in the terms. The Register’s reporting put a rough number on the reach: at least a million businesses already use Meta’s AI agents inside WhatsApp and Messenger, and nearly every news publisher operates at least one Meta business account, which puts both groups squarely inside the same clause.
The timing lines up with how much weight Meta’s own platforms already carry in AI-generated answers. A related pattern on this site found that Facebook content gets cited roughly 250 times more often than TikTok’s inside AI Overviews, a gap this crawling clause could widen further by feeding more of that content straight into Meta’s own pipelines.
Compare that with the Commercial Terms dated March 6, 2024: the crawling paragraph did not exist. That earlier document covered only content a business had itself posted inside Meta’s own products, the right to display and distribute it there. The crawling and test-credentials language is new to the general Commercial Terms; before this update, it lived only in the Self-Serve Ad Terms, scoped to businesses that bought ads.
The purpose wording moved in four days, but not everywhere
Search Engine Journal first reported the broadened clause on October 16, 2026. The same day, Meta communications director Andy Stone posted publicly that nothing had changed in how the company uses web crawlers and that businesses could still rely on robots.txt to set their preferences. By October 17, outlets including The Verge found that Meta had edited the US English preview of the document: the word “train” disappeared from the list of purposes, and the scope narrowed from the broadly worded “Meta Company Products, services and features” to “advertising and business products” specifically.
Checking Meta’s own Commercial Terms page directly this week shows the edit did not reach every edition. The US English version now reads that Meta may use crawled content “to provide, develop, improve, research, and maintain advertising and business products.” The French-language preview, read the same day, still carries the original sentence in full: Meta may “fournir, développer, améliorer, rechercher, entraîner et maintenir l’intégrité des Produits, services et fonctionnalités des Entités Meta,” entraîner being the French for train. PPC Land reported the same split on October 18, finding the UK English edition unchanged as well. The crawling authorization itself, the right to use automated tools and test credentials against a business’s domains, reads identically in every edition checked. Only the stated purpose text moved, and only in some languages, ten days before the terms bind every business account on the platform. PPC Land’s review of the narrower wording found the same open question several legal observers raised about the original version: nothing in the document defines what counts as a “business product,” and the phrase “associated with your products or services” is broad enough that a site’s entire domain, not just a product page, could qualify.

Robots.txt is a promise outside the contract, not a clause inside it
Andy Stone’s reassurance rests on robots.txt, the text file site owners use to tell automated crawlers which pages to skip. Eric Goldman, a law professor at Santa Clara University, told The Register that the terms “don’t expressly say that Meta will respect those robots.txt instructions,” and that nothing in the law compels a company to honor them either. Following robots.txt remains a goodwill gesture, not an enforceable commitment. That distinction sits at the center of the governance question for any business reviewing its Meta presence before October 30: the contract grants the crawling right outright, and the only promise to limit it lives in a spokesperson’s public post, not in the document a business agrees to by continuing to use the product.
Meta’s own crawler documentation adds a second complication. The tool behind this activity identifies itself as meta-externalagent, and Meta’s webmaster guide describes it as the crawler used both to train its foundation AI models and to index content directly for product features. A single robots.txt entry controls that one user agent. A business can disallow meta-externalagent entirely, which blocks both uses at once, but nothing in Meta’s own documentation offers a way to keep a page indexed in Meta’s products while opting the same content out of AI training. The two uses travel on the same switch, which is also the dilemma behind who stays accountable when an automated system acts on a business’s data without a person reviewing each step.

Worth separating clearly from all of this: the Commercial Terms are a contract, not a regulation. They sit apart from the EU’s AI Act, which this site has already covered for its own timeline of obligations; nothing here changes what that regulation requires, and nothing in the AI Act currently reaches a platform’s own crawling clause.
None of this makes the clause illegal, or the crawling itself a trap set specifically for small business owners. Meta is entitled to propose the terms it proposes, and continued use of a free business tool has generally come with some kind of exchange attached. What changed this month is how visible that exchange became, and how unstable its exact wording turned out to be in the days before it takes effect. If you manage a Facebook Page, an Instagram account, or a WhatsApp Business catalog, you have three things worth checking before October 30: which language edition of the Commercial Terms actually governs your account, what your own robots.txt file currently allows for meta-externalagent, and whether you are comfortable with indexing and training staying linked for now. None of those checks require a lawyer, just a few minutes on a page you have probably never opened.
Photo credits: Pixabay.








Pingback: Meta Pauses New Business Terms: Crawl Rights Still Live