Guides · 2026-10-01

Claude account disabled: Anthropic’s stated reasons, the appeal route, and what not to do

The short version: if your Claude account is disabled, Anthropic’s route is to sign in to claude.ai with that account and submit an appeal from the restricted page, which its Safeguards team then reviews. There is no promised outcome, and the Usage Policy explicitly prohibits getting around a ban with a new account. Before appealing it helps to know which message you’re actually seeing and which region Claude sees your connection coming from, but a checker can’t tell you why an account was disabled, and it can’t get one back.

Sources, all read on 2026-10-01: Anthropic’s Help Center article Safeguards warnings and appeals, the Supported Regions Policy, the Usage Policy, the Consumer Terms of Service and Claude Code’s error reference. User reports are labeled as such.

Which message are you seeing?

Being locked out can come from the account, an organization, a credential or the exit region, and each is handled differently:

• Account disabled on the web: the text users have posted in Anthropic’s GitHub repository is Your account has been disabled after an automatic review of your recent activities. (user report). Others have posted a suspension email saying an internal investigation of “suspicious signals” found a Usage Policy violation (user report).
• Account on hold in Claude Code: the error reference gives Your account is on hold and can't use Claude Code. View details or appeal: https://claude.ai/restricted. Signing in again with the same account doesn’t clear it, because the hold is on the account, not the login.
• This organization has been disabled.: in Claude Code this appears as API Error: 400 ... This organization has been disabled. The documented cause is a stale ANTHROPIC_API_KEY from a disabled Console organization. Environment variables take precedence over /login, so the old key wins even if you have a working Pro or Max subscription. Run /status to see which credential is active; if no key is set and the error persists, contact support. Users have also reported seeing the same text on claude.ai (user report).
• Region-not-supported messages: these mean that request was refused. On their own they are not a notice that the account is disabled.

The reasons Anthropic lists

The Help Center’s appeals article names three:

• Repeated Usage Policy violations: the Usage Policy (effective 2025-09-15) says Anthropic may throttle, suspend or terminate access when it learns of a violation.
• Account creation from an unsupported location: neither the Help Center list nor the API and Claude.ai lists on the Supported Regions Policy page include mainland China, Hong Kong or Macau. Taiwan is listed.
• Terms of Service violations: the Consumer Terms (effective 2025-10-08) allow use only in line with the Usage Policy and the Supported Regions Policy, and let Anthropic suspend or terminate access without notice if it believes you’ve breached them.

Age is covered separately. The terms require you to be at least 18, or the minimum age to consent in your location if that’s higher. Per the Help Center’s age assurance article, accounts showing indicators of a user under 18 are disabled; the notification email links to age verification through Yoti, and passing it reinstates the account.

For scale, Anthropic’s transparency report lists 11.4 million banned accounts, 398k appeals and 42k appeal overturns for January to June 2026. It also names Supported Region Policy violations among the grounds for warnings, suspensions and terminations.

How to appeal

1. Sign in to claude.ai with the disabled account. The Help Center says the appeal form is only available while signed in; it’s the Submit an appeal link at claude.ai/restricted.
2. If your own account is fine but an organization you belong to was paused for unusual activity, the restricted screen lists each organization on hold. Click Request a review on the affected one to ask the Safeguards team to look again.
3. Then wait. Both the Help Center and the appeal form note that responses are slower than usual because of high email volume.

Anthropic doesn’t publicly describe the fields of the in-app form, so this article doesn’t either. The Help Center’s identity verification article links a separate appeal form (a Google Form titled Account Ban Appeal) with three fields: name and the email associated with your account (both required), plus an optional message. Its intro says to fill it in if you believe your account was wrongly disabled by Anthropic’s systems.

If you only received a warning and think it was a mistake, the Help Center says to email [email protected] with the details and your account information. Every banned user can sign in and delete their own account, and self-serve data export is available too, though what you can export may be restricted depending on the violation.

What the terms don’t allow

The Usage Policy section on not abusing the platform includes, among others:

• Circumventing a ban with a different account, whether by creating a new one, using an existing one, or giving access to a person or entity that was banned.
• Giving people, entities or users account or API access to Claude in violation of the Supported Regions Policy.
• Coordinating violations across multiple accounts to avoid detection, or automating account creation.

The Consumer Terms also bar sharing your account, login details or API key, and bypassing Anthropic’s systems or protective measures. And a Help Center note says that if an account was disabled for Usage Policy violations, its email address can’t be used for a new account, however long you wait.

Check the network side before you appeal

One of the listed reasons is about location, and what Claude sees directly is the IP your connection exits from. Anthropic doesn’t publish how it determines location or how much weight an IP carries in a ban decision. What you can confirm from outside is the exit Claude’s side sees:

1. Open the Claude IP check. The Claude row in the AI card is the exit your browser gets when it asks claude.ai directly, marked with whether that region is on the supported list, the IP type, and any leaks. With per-domain proxy rules this exit can differ from what other sites see; see Which IP do ChatGPT and Claude actually see?
2. If you use Claude Code, run curl -sL ipjudge.org/en/cc | bash in the same terminal to see the exit api.anthropic.com sees (details on the terminal exit check page).
3. An exit in mainland China, Hong Kong or Macau is outside Anthropic’s lists. If you are actually located there, you’re outside the service area under Anthropic’s own policy, and a different exit doesn’t change that.

The appeal form has a message field where you can state where you actually are, what network you use and when the message appeared. Anthropic doesn’t say how such details affect the outcome. How IP Judge reaches its verdicts is on the method page.

FAQ

How long does an appeal take?

Anthropic gives no timeline. The Help Center and the appeal form only say responses are currently slower than usual.

Will an appeal get my account back?

Not necessarily. The transparency report shows 42k overturns out of 398k appeals in the first half of 2026, roughly one in ten.

Can I sign up again with another email?

The Usage Policy prohibits circumventing a ban with a new or different account, and an email tied to an account disabled for Usage Policy violations can’t be used to sign up again.

I’m signed in but can’t find the appeal form. What now?

In June and July 2026 users reported that claude.ai/restricted redirected to the home page and only showed This organization has been disabled.; a later comment said it was working again (user report). The Help Center’s identity verification article also links an appeal form you can use.

Will I get a refund for my subscription?

The Consumer Terms say there’s no refund if access is terminated for a violation of the terms. If Anthropic ends a subscription for any other reason and you bought it on its website, it refunds the unused portion pro rata.