Claude’s region rules take effect November 12. Who they cover, and why it asks for ID

What takes effect on November 12

Anthropic’s Supported Regions Policy now reads “Effective Nov 12, 2026,” the day its new Usage Policy takes effect too. It opens: “Anthropic’s products and services — including our apps (such as Claude.ai, Claude Code, and Cowork), our API and developer platform, and Claude models offered through cloud platforms and authorized resellers — are available only in the countries and regions listed below.”

The list has 185 entries. “Any country or region not listed is unsupported,” and the policy names three ways to fall outside it:

Who	The policy’s words	For example
People in an unsupported region	“Use by persons while physically located in an unsupported region, including users and personnel of otherwise supported entities.”	A US employee on a trip to Hong Kong
Companies based in one	“Use by entities incorporated or headquartered in an unsupported region and their users or personnel, regardless of whether such individuals are physically located in a supported region.”	A Hong Kong company’s staff in London
Companies controlled from one	“Use by entities that are majority-owned or controlled, directly or indirectly, by persons or entities in unsupported regions, including use by their users or personnel, regardless of whether such individuals are physically located in a supported region.”	A Singapore subsidiary of a company in mainland China

China, Hong Kong and Macau aren’t on the list. Neither are Russia, Belarus, Iran, North Korea, Cuba, Syria, Venezuela, Afghanistan, Myanmar or Yemen, and Ukraine is listed “except Crimea, Donetsk, Kherson, Luhansk, and Zaporizhzhia regions.” Taiwan is on it. “Territories and dependencies of a listed country are also supported (for example, Puerto Rico as a territory of the United States).”

The Usage Policy carries the rule into what you may do with Claude. It forbids using Anthropic’s services to “Access, facilitate, or provide others (including end users) with access to our products or services in violation of our Supported Regions Policy,” and to “Resell, proxy, or otherwise provide access to Claude through unauthorized means, including services that route requests through consumer subscriptions or misrepresent the product or client being used.” For a suspected violation, Anthropic “may warn you or throttle, limit, suspend, or terminate your access.”

What changed, and what didn’t

Anthropic calls it a clarification. Its October 8 post on the new Usage Policy says: “We’ve updated our Supported Regions page to clarify how we enforce this policy. Our policy prohibits the use of Claude by people who are physically located in an unsupported region, by entities who are incorporated or headquartered in one, and by entities who are majority-owned or controlled by persons or entities in those regions.”

The old page said less. On September 2, 2026, as the Wayback Machine kept it, it had separate lists for “commercial API access” and “Claude.ai access,” and one sentence about owners: “To the extent permitted by law, Anthropic reserves the right to not provide its products or services to entities whose majority direct or indirect ownership is attributable to nations other than those listed in our Supported Regions Policy.” That came from Anthropic’s September 2025 change, which reached entities “more than 50% owned, directly or indirectly, by companies headquartered in unsupported regions.”

What’s new is in the wording:

1. One list for everything, naming Claude Code, Cowork, cloud platforms and authorized resellers alongside Claude.ai and the API.
2. The page spells out that where you are counts, even for “users and personnel of otherwise supported entities.” A trip is enough.
3. “Controlled” sits beside “majority-owned,” and “reserves the right” is gone.
4. The Democratic Republic of the Congo is on the list, which it wasn’t in September.

What didn’t change is the list itself, apart from that: no country or region came off it.

The new text was on the page by October 13, under an effective date of May 12, 2026; the page now gives November 12.

The ID check, and the reports of bans

Anthropic’s help article, Identity verification on Claude, dated August 11, 2026, says: “We are rolling out identity and business verification for a few use cases, and you might see a verification prompt when accessing certain capabilities, as part of our routine platform integrity checks, or other safety and compliance measures.” The check runs through Persona, and it asks for:

• “A valid government-issued photo ID: the physical document, in hand”: a passport, driver’s license, state or provincial ID card, or national identity card, unexpired and with your photo.
• A phone with a camera, because “you may be asked to take a live selfie.”
• For a business, its “legal name, address, and business registration number.”

It won’t take “Photocopies, screenshots, scans, or photos of a photo,” “Digital or mobile IDs,” non-government IDs such as student IDs and employee badges, or temporary paper IDs. “Your ID and selfie are collected and held by Persona, not on Anthropic’s systems,” and Anthropic says “We are not using your identity data to train our models.”

Anthropic hasn’t said what triggers a prompt. Its Privacy Policy lists what it collects from your device, including your “time zone setting, IP address (including information about the location of the device derived from your IP address).”

The reports so far come from users. On November 12, MarsBit, a Chinese crypto and tech news site, reported that from the day before, many Claude users had been asked to verify and then banned, including paid accounts used for a year or more, some in the middle of a Claude Code task. By its account, a Max 5x subscriber passed with a Chinese passport and was banned at once; another user who had passed with a Chinese passport on October 31 was asked again on November 12 and banned; and some who passed with Chinese passports still had access. It also reported people stuck in the flow, unable to change a wrongly chosen document type when asked for a second ID, and a help page that kept redirecting. TokenPost, a Korean outlet, wrote the same day that “When an account shows signs of activity from an unsupported region, Anthropic may request identity verification. Accounts that fail to verify are banned,” without saying where that came from. Anthropic’s own pages don’t say it in those words.

It isn’t the first wave. On October 2, the South China Morning Post reported that Anthropic “appears to have stepped up enforcement of geographical restrictions on its Claude models in Hong Kong, with local users reporting abrupt account suspensions after accessing the service through virtual private networks (VPNs).”

Passing the check doesn’t settle it. The help article asks itself “Why did my account get banned after verification?” and answers: “As part of our safety processes, we may ban an account for a variety of reasons, including repeated violations of our Usage Policy or Terms of Service.”

Who it affects

• You live in a supported country and use Claude there: the policy asks nothing new of you.
• You travel to mainland China, Hong Kong or Macau: the policy covers you while you’re there, on a personal account or your employer’s. Plan the trip’s work without Claude.
• Your employer is incorporated or headquartered in an unsupported region, or majority-owned or controlled from one: the policy covers all its staff, wherever they sit, and every route to Claude, since it names cloud platforms and authorized resellers too.
• You buy Claude through a reseller or a cheap proxy: the Usage Policy forbids anyone to “Resell, proxy, or otherwise provide access to Claude through unauthorized means,” so an unauthorized one is breaking Anthropic’s rules. Anthropic’s September threat report also found a reseller “offering cheap Claude access—which turned out to be neither cheap nor actually Claude,” and advises that “AI access should be purchased only through authorized channels.”
• You hold a Chinese passport but live in a supported country: the policy doesn’t mention nationality or passports. It turns on where you are and who your employer is.

What to do if Claude asks for ID, or bans you

1. Verify only from inside Claude. Start from claude.ai or the app rather than a link in an email. Anthropic’s threat report describes “lookalike verification domains whose reverse proxy relayed the real know your customer (KYC) flow” while the operator “captured the verified session and documents.”
2. Have the physical ID and your phone ready, and retry in better light if it fails: “You’ll have multiple attempts within the verification flow.” Out of attempts, use Anthropic’s help form.
3. If you’re banned, appeal from the banned account. Anthropic’s appeals article says to log in at claude.ai, where the form sits beside options to export your data or delete the account: “You don’t need to contact Anthropic to do this.” For a warning you think was a mistake, email usersafety@anthropic.com.
4. Don’t count on a refund. Anthropic’s Consumer Terms say “If we terminate your access to the Services due to a violation of these Terms and you have a Subscription, you will not be entitled to any refund.” If you subscribed through an app store, its terms apply.
5. If you can’t use Claude where you work, an open-weight model on your own hardware is outside Anthropic’s policy, though each comes with its own license. Our guide to local coding agents compares the ways, and Kimi K3.1 Pro and Mistral Large 4 are recent ones with weights to download.

What Anthropic hasn’t said

• What triggers a verification prompt, and whether November 12 started a sweep.
• How many accounts it has asked to verify or banned. MarsBit noted the same gap.
• Why some accounts were banned right after passing, and others weren’t.
• How it decides who “controls” a company, a word the policy doesn’t define.
• Whether a short trip or a layover counts. The policy’s text has no exception for either.

We’ll update this memo if Anthropic says more.