Skip to content

Claude Ban Causes: What Anthropic Documents, What It Doesn’t

Anthropic documents three ban reasons. The causes people blame most — datacenter IPs, VPNs, time zones — are not among them.

See the AI-subscription ISP
Pricing

The three reasons Anthropic documents

Anthropic’s help center says it may ban an account for repeated violations of its Usage Policy, for creating the account from an unsupported location, and for violations of its Terms of Service. For prompts that break the Usage Policy it warns first; for API customers the warnings are tied to ongoing thresholds of violative behavior across the whole API account.

The scale is public. Anthropic’s Transparency Hub shows 11.4 million banned accounts, 398 thousand appeals and 42 thousand overturned bans for January–June 2026 (page updated 23 July 2026). Divided, that is about one appeal in ten overturned — our arithmetic, not an Anthropic figure. The numbers count accounts, not people, and the page gives no breakdown by reason, so nobody can tell from it how many bans had anything to do with an IP address.

Cause by cause: how well it is documented

Cause by cause: how well it is documented
ItemWhat triggers itDocumented by Anthropic?What you can do
Usage Policy violationsWhat you ask Claude to do, repeated after warningsYes: help center and Usage PolicyRead the warning and change the prompts; appeal if it was a mistake
Unsupported locationCreating or using an account while physically in an unsupported region, or being a company owned or controlled from oneYes: Supported Regions Policy and the phone-verification articleUse Claude only from a supported region where you physically are. An IP address does not change where you are
Shared accounts and loginsSharing your login or API key, or letting someone else use your accountYes: Consumer Terms, section 2One account per person (details below)
Another account after a banOpening a new account, or using someone else’s, to get around a banYes: Usage PolicyAppeal instead of opening a new account
PaymentStolen, fraudulent or unauthorized payment methodsYes: Usage PolicyPay with your own card, in your own name
Card declinedA billing address that does not match the bank’s record or the card’s countryDocumented as a cause of declined payments, not as a ban reason: card-declined articleKeep the billing address your bank has on file (see the environment guide)
Reselling accessReselling or proxying access to Claude through unauthorized meansYes: Usage PolicyDo not sell or pool your subscription
Datacenter, VPN or shared IPThe type or reputation of the address you connect fromNo: not listed as a ban reasonTreat it as a habit: one account on one stable address that nobody else uses
Time zone, language, deviceSettings that contradict your IP or your billing countryNoKeep them consistent; it costs nothing

Documented means stated in the Anthropic pages we read on 29 October 2026. “No” means we found no statement either way. It does not mean those signals are ignored: only Anthropic knows what its systems use.

Shared accounts: the cause you control

The Consumer Terms say you may not share your account login information, your Anthropic API key or your account credentials with anyone else, and may not make your account available to anyone else. That covers a Pro or Max plan split between colleagues, a family login used from several homes, and a password handed to a freelancer. Give every person their own account.

The same reasoning applies to buying an account from someone else. We found nothing in Anthropic’s pages that supports it, and the seller can still hold the recovery email and phone number. Phone numbers add a limit of their own: one number can verify at most three Claude accounts, and a verified number cannot be changed.

Sharing also shows up on the network side. Several people on one login means several addresses, devices and time zones for one account, and several accounts on one address means that address carries everyone’s behavior. The IP guide explains why one account should have one address that nobody else uses.

An error is not always a ban

“API Error: 403 Request not allowed” after you log in to Claude Code. Claude Code’s own troubleshooting page asks you to check that your Pro or Max subscription is active, that a Console account has the Claude Code or Developer role, and that no corporate proxy is interfering with API requests. The help center describes a restricted screen for banned accounts, from which you reach the appeal form after logging in at claude.ai.

“This organization has been disabled” although you pay for a subscription. The Claude Code documentation says an ANTHROPIC_API_KEY environment variable, often left over from an old project or employer, overrides the subscription login. Remove the variable before you assume the account is banned.

A 403 when you test your network from the terminal is, according to the same page, usually a proxy or network filter blocking the host, or Claude Code not being available in your region. A proxy cannot make an unsupported region supported.

On Claude Enterprise, “access denied” can come from the organization’s IP allowlist: requests from addresses outside it are blocked. Ask your administrator before you appeal.

What nobody outside Anthropic can confirm

Anthropic’s ban article does not mention VPNs or IP addresses, and neither does its Supported Regions Policy, which speaks of people “physically located” in a region. Many articles about Claude bans name datacenter IPs, frequent IP changes or browser fingerprints as causes. The ones we read give no thresholds — “frequent” and “short time” are never defined — and rest on one user’s report or on community guesses; several are written by sellers of proxies, servers or relay services. We therefore give you no number and no promise.

What we can do is remove the contradictions that are under your control: one account per person, one stable address per account, a time zone, language and billing country that agree with it, and one payment method that is yours. Platform risk-control policies change from time to time, and we cannot guarantee that any account will stay 100% free of bans.

If your account is already banned, the way back is the appeal form, not a new address.

Claude ban causes FAQ

Does Claude ban accounts for using a VPN or a proxy?

Anthropic’s ban article and Supported Regions Policy do not mention VPNs, proxies or IP addresses. The Usage Policy does forbid reselling, proxying or otherwise providing access to Claude through unauthorized means, which is about providing access to others, not about how you connect.

Is a datacenter IP a documented ban reason?

No. Datacenter, VPN and shared addresses are not listed among the three reasons. We cover how IP databases label them, and why a dedicated address is the safer habit, in the IP guide.

Can two people share one Claude Pro account if they use the same IP?

No. The Consumer Terms forbid sharing your login or making your account available to anyone else, whatever address you connect from.

How many Claude accounts can I have?

We found no account cap in the pages we read. One phone number can verify at most three accounts, and the Usage Policy forbids creating multiple accounts to avoid detection.

How common are bans, and how many are reversed?

For January–June 2026 Anthropic reports 11.4 million banned accounts, 398 thousand appeals and 42 thousand overturned bans. Those figures count accounts and carry no breakdown by reason.

Will changing my IP get a banned account back?

No. Only an appeal can lift a ban, and opening another account to keep using Claude is itself a Usage Policy violation.

Reference links

9
  1. Claude Help Center: Safeguards warnings and appeals

    Official documentation · Confirmed on:

    Read on 29 October 2026. Ban reasons listed: repeated violations of the Usage Policy, account creation from an unsupported location, Terms of Service violations. Users are warned when Anthropic believes their prompts break the policy; for API customers the warnings are tied to thresholds across the whole API account. The appeal form opens only after you log in at claude.ai with the banned account; banned Free, Pro and Max users can still export their data (possibly restricted, depending on the violation) or delete the account. The article does not mention VPNs or IP addresses and says response times are longer than normal.

  2. Anthropic Transparency Hub: platform security and enforcement figures

    Official documentation · Confirmed on:

    Read on 29 October 2026. For January–June 2026 (page updated 23 July 2026): 11.4 million banned accounts, 398 thousand appeals, 42 thousand appeal overturns. The page does not define “ban” beyond saying Anthropic may warn, suspend or terminate access, and gives no breakdown by reason. The ratio of about one appeal in ten is our own division of 42 by 398.

    The figures count accounts, not people.

  3. Anthropic: Supported Regions Policy

    Official documentation · Confirmed on:

    Read on 29 October 2026: use by persons while physically located in an unsupported region is not supported; the policy also covers entities majority-owned or controlled by persons or entities in unsupported regions. The page does not mention VPNs, proxies or IP addresses.

    The page shows an effective date of 12 November 2026.

  4. Anthropic: Consumer Terms of Service (effective 8 October 2025)

    Official documentation · Confirmed on:

    Read on 29 October 2026: section 2 — you may not share your account login information, Anthropic API key or account credentials with anyone else, and may not make your account available to anyone else; you must be at least 18 or the minimum age to consent in your location, whichever is higher. Section 12 — if access is terminated for a violation of the Terms, a subscriber is not entitled to a refund.

  5. Anthropic: Usage Policy

    Official documentation · Confirmed on:

    Read on 29 October 2026, account and platform integrity section: no using a different account, such as a newly created one, to continue after a ban, and no giving access to a person who was banned; no coordinating malicious activity across multiple accounts or creating multiple accounts to avoid detection; no automation in account creation; no stolen, fraudulent or unauthorized payment methods; no reselling, proxying or otherwise providing access to Claude through unauthorized means.

    The version on the site shows an effective date of 12 November 2026.

  6. Claude Help Center: Why was my card declined?

    Official documentation · Confirmed on:

    Read on 29 October 2026: the billing address on the payment method must match the address your bank has on file, and even small differences such as a missing accent or a misspelled street name can cause a decline; check the list of supported billing locations; the billing address must match the payment method’s country of origin. The article does not mention VPNs or IP-based location checks.

  7. Claude Help Center: Verify your phone number

    Official documentation · Confirmed on:

    Read on 29 October 2026: only users physically located in a supported location can create and use Claude accounts; VoIP numbers, Google Voice, numbers created in apps and landlines cannot be used; one number can verify at most three Claude accounts; a verified number cannot be changed.

  8. Claude Code docs: Troubleshoot installation and login

    Official documentation · Confirmed on:

    Read on 29 October 2026. “403 Forbidden after login” (API Error: 403 Request not allowed): verify a Pro or Max subscription is active; for Console users confirm the Claude Code or Developer role; behind a proxy, corporate proxies can interfere with API requests. In the network check, a 403 is “usually a proxy or network filter blocking the host, or Claude Code is not available in your region”. “This organization has been disabled” despite an active subscription: an ANTHROPIC_API_KEY environment variable overrides the subscription login.

  9. Claude Help Center: Restrict access to Claude with IP allowlisting

    Official documentation · Confirmed on:

    Read on 29 October 2026: IP allowlisting is available on Enterprise plans only; requests from IP addresses not on the allowlist are blocked and users see an access-denied message; administrators compile the CIDR ranges to allow and send them to Anthropic, and omitting a required range can lock users out.

Ready to try it?

Create an account, top up once, then pick rotating residential, static ISP or datacenter IPs and have one running in your browser within minutes.

Get started