Account Safety Guide

Why Claude Accounts Get Banned: Appeals, Refunds and Compliant Access

As of 2026-11-18 (previous checks: 2026-11-12 and 2026-11-09), Anthropic's Supported Regions Policy (effective 2026-11-12 at anthropic.com/legal/supported-regions, enforcing three simultaneous dimensions: physical location, entity incorporation or headquarters, and >50% ownership or control) and updated Usage Policy (effective 2026-11-12 at anthropic.com/legal/aup) state that "If we suspect that you may have violated our Usage Policy", Anthropic may warn you or throttle, limit, suspend or terminate your access, strictly prohibit creating new accounts to circumvent a ban, and note that a real-time safeguard blocking an output does not by itself indicate a violation. The Help Center (14328960) explains that identity verification is handled by Persona (requiring a physical government photo ID and a live selfie) and lists three reasons an account can be banned: repeated Usage Policy violations, account creation from an unsupported location and Terms of Service violations; in Claude Code, restricted accounts trigger exact error strings such as "Your account is on hold and can't use Claude Code. View details or appeal: https://claude.ai/restricted", "This organization has been disabled", or "OAuth token revoked". After a ban, log in to claude.ai with the banned account and submit Anthropic's official appeal form at https://claude.ai/restricted (the Help Center notes that response times are currently longer than normal); refund requests go through Anthropic's official refund process, but under the Consumer Terms a subscriber whose access is terminated for violating the terms is not entitled to a refund. A QCode API key is an independent service, distinct from a claude.ai subscription, and each is governed by its own terms. Anthropic disables accounts after automatic reviews for reasons many users never see coming. Here is a calm, factual look at the triggers, the appeal and refund steps, and how to choose per-token API access for team and automation work.

One key for Claude, GPT and leading Chinese models, billed per token, up and running in minutes.

QCode.cc is an independent API access service, not affiliated with or endorsed by Anthropic; product names are trademarks of their owners, and this information is general reference only, so please consult the official terms.

Why Claude Accounts Get Disabled

The ban reasons below relate to the Claude Help Center, the Consumer Terms and the Usage Policy; knowing them helps you stay within the terms. According to public posts by two authors on Habr (2026-10-16 and 2026-11-09), a wave of bans came around 1 October 2026 and a new wave of ban complaints followed on 28–29 October (the authors' own accounts, not Anthropic data).

1

Third-party tools using subscription tokens

Feeding your Claude Pro or Max OAuth token into tools like OpenCode, Cline, Roo Code or Goose violates the Consumer Terms; since January 2026 Anthropic actively blocks this, and from April 2026 subscriptions no longer cover third-party-tool usage.

2

Unsupported region or location

Claude is unavailable in mainland China, Russia, Iran and other unsupported regions; accounts created or used from these locations can be disabled, and since September 2025 Chinese-controlled companies are prohibited from using the service. As of 2026-11-18, the Supported Regions Policy effective 2026-11-12 (anthropic.com/legal/supported-regions, covering Claude.ai, Claude Code, Cowork, the API and developer platform, and cloud or reseller channels) enforces three simultaneous dimensions: (1) individuals physically located in an unsupported region (including employees of supported-region companies traveling in an unsupported region); (2) entities incorporated or headquartered in an unsupported region and their personnel anywhere; and (3) entities directly or indirectly owned (>50%) or controlled by persons or entities in unsupported regions (mainland China, Hong Kong, Macau, Russia, Iran, Belarus and others are not on the supported list, whereas Germany, the UAE, Saudi Arabia, Qatar, Egypt and Taiwan are on the supported list).

3

Repeated Usage Policy violations

The Claude Help Center lists repeated Usage Policy violations as a ban reason, and Anthropic warns users when it believes their prompts violate the Usage Policy.

4

Terms of Service violations

The Claude Help Center lists Terms of Service violations as a ban reason, and the Consumer Terms state that Anthropic may suspend or terminate access without notice if it believes you have breached them.

5

Account sharing, reselling and limit evasion

Sharing or reselling accounts, plugging subscription tokens into third-party tools and evading usage limits all lead to bans, and the Usage Policy explicitly prohibits circumventing a ban through another account, new or existing. As of 2026-11-18, the updated Usage Policy effective 2026-11-12 (anthropic.com/legal/aup) sets the enforcement threshold at "If we suspect that you may have violated our Usage Policy" (suspicion alone can trigger warnings, throttling, suspension or termination), prohibits creating new accounts to circumvent a ban, and forbids unauthorized reselling or routing requests through proxied personal subscriptions.

How to Stay Compliant

A few habits keep your usage inside the Consumer Terms; each follows Anthropic's own guidance.

Use the official Claude Code CLI on your own machine or an Anthropic Console API key, both of which are permitted.

Keep the account for your own use only, never share or resell it, and follow Anthropic's Usage Policy.

Never route your Pro or Max subscription OAuth token through third-party tools like OpenCode, Cline or Goose.

For production, automation or team workloads, use pay-as-you-go API keys, which is the path Anthropic itself directs this usage to.

Subscription vs Pay-as-you-go API

Both have a place; this honest comparison shows why the API is the stable path for automation and teams.

Dimension Subscription (Pro/Max) Pay-as-you-go API
Ban / suspension risk Personal account can be disabled by an automatic review. Governed by API terms designed for programmatic and production use.
Third-party tools Routing subscription tokens through them violates the terms. Third-party tools are permitted when used with a Console API key.
Region availability Limited to supported consumer regions for sign-up and use. Still subject to the vendor's region policy; check the officially supported regions.
Automation and production use Subscriptions are intended for interactive individual use, not sustained scripted or automated production workloads. Built for scripted and automated production workloads.
Cost predictability Fixed monthly fee, but usage caps and review risk apply. Per-token billing you can watch in real time.

The Stable Path: API-based Access

For automation or team work, API keys billed per token match the vendors' own guidance; API access services such as QCode.cc are one example of this approach.

No reliance on a personal subscription login

You use an API key billed per token rather than a personal subscription login, and Anthropic's Usage Policy applies as usual.

One key for Claude, GPT and Chinese models

A single API key reaches Claude, GPT or Codex and Chinese models through their native protocols, with no juggling of separate logins.

Per-token billing you can watch

Usage is billed per token, with per-model rates listed on /models, and you can look up the model and usage of each request at probe.qcode.cc with your own key instead of guessing against a flat cap.

Fully self-service, up in minutes

Fully self-service: top up, activate instantly, and start using right away with no manual processing, plus fair upgrades whenever you need more.

View plans

Frequently asked questions

Why was my Claude account suspended or banned?

As of 2026-11-12, the Claude Help Center lists three ban reasons: repeated Usage Policy violations, account creation from an unsupported location, and Terms of Service violations. Warnings are part of the same safety process: the Help Center says it warns users when it believes their prompts violate the Usage Policy. Plugging a subscription OAuth token into third-party tools, sharing or reselling accounts, and evading usage limits are the common violations covered above. After a ban, start with the official appeal process described below.

Can using Claude Code, or prompting in a language other than English, get my account banned?

Neither using Claude Code nor the language of your prompts, Chinese included, is among the ban reasons the Help Center lists, which are repeated Usage Policy violations, account creation from an unsupported location, and Terms of Service violations; whether an account is banned is for Anthropic's review to decide. Running the official Claude Code CLI on your own machine is allowed and is built for scripted and automated use. What violates the Consumer Terms is putting your Pro or Max subscription OAuth token into a different client, such as OpenCode, Cline or Goose. Third-party tools are fine when they use an Anthropic Console API key instead.

What should I do if my Claude Code account is banned? Can I appeal?

Appeal first: log in to claude.ai with the banned account and open the official appeal form (the Help Center notes that you must be logged in to access it). You can appeal, but success is far from guaranteed. According to Anthropic's Transparency Hub, about 1,700 of roughly 52,000 appeals succeeded in the second half of 2025; in the first half of 2026 (the page is marked last updated 2026-07-23) about 11.4 million accounts were banned and about 42,000 of roughly 398,000 appeals were overturned. The outcome is up to Anthropic's review; while you wait, you can log in to claude.ai and export your data.

Can using Claude from an unsupported region get an account banned?

It can. One of the three ban reasons in the Claude Help Center is account creation from an unsupported location, and as of 2026-10-15 Anthropic's supported-regions list does not include mainland China. Whether an account may be used is governed by Anthropic's policies. As of 2026-11-18, the Supported Regions Policy effective 2026-11-12 further enforces three simultaneous dimensions (physical location, entity incorporation or headquarters, and >50% ownership or control); the Help Center (14328960) also states that identity verification is conducted via Persona using a physical government photo ID and a live selfie, and completing verification does not guarantee an account will not be banned for violation signals.

Which way of using Claude models fits production, automation or team work?

For production, automation or team use, API keys billed per token are the vendor-recommended path, since Anthropic itself directs this usage to the API; billing is per token and usage is visible. QCode.cc is one API-based access service that reaches Claude, GPT and Chinese models through a single key, with per-model rates on /models and quota packs on /pay-as-you-go. Anthropic's Usage Policy applies to this use as well.

After a ban: appeals, refunds, your data and access options

First step after a ban: the official appeal

As of 2026-11-18 (the same when checked on 2026-11-09 and 2026-10-15), the Claude Help Center lists three reasons an account may be banned: repeated violations of the Usage Policy, account creation from an unsupported location, and Terms of Service violations. Per the official Claude Code error reference (code.claude.com/docs/en/errors, 2026-11-16), exact error strings when an account or organization is restricted include "Your account is on hold and can't use Claude Code. View details or appeal: https://claude.ai/restricted", "This organization has been disabled", and "OAuth token revoked" or "Your account does not have access to Claude.". In addition, the Help Center (14328960) notes that identity verification is handled by Persona, requiring a physical government photo ID and a live selfie, and passing verification does not guarantee immunity from bans. If you believe the ban is a mistake, the official route is to log in with the restricted account and submit the appeal form at https://claude.ai/restricted (or email usersafety@anthropic.com if you only received a warning).

You fill in the appeal after logging in to claude.ai with the banned account; the Help Center says 'You must be logged in to access the appeal form', and when checked on 2026-10-15 and again on 2026-11-09 it noted that response times were longer than normal (it gives a recent launch and higher email volume as the reason). Anthropic's Consumer Terms state that it may suspend or terminate access without notice if it believes you breached the terms, so an appeal has no guaranteed outcome. If your own account is in good standing but an organization you belong to has been paused, the restricted-account screen lists each organization on hold, and you can click "Request a review" to ask Anthropic's Safeguards team to take another look. As of 2026-11-12, the Help Center says the appeal goes to Anthropic's Safeguards team, which can further investigate why your account was disabled; if you believe a warning was issued by mistake, you can email the details of your situation and your account information, using the address given on that page.

  • Log in to claude.ai with the banned account and open the appeal form.
  • Explain why you believe the account was disabled by mistake.
  • From the same screen you can export your data first, or delete the account yourself.

Can you get a refund after a ban?

Usually not. Anthropic's Consumer Terms (version effective 2025-10-08) say that if access is terminated for a violation of the terms and you have a subscription, you will not be entitled to any refund; if Anthropic ends a subscription bought on its website for any other reason, it refunds the remaining portion pro rata.

The Help Center's refund page adds that, except as expressly provided in the Consumer Terms, where required by law, or in cases of fraud, all payments are non-refundable; for a payment the terms define as refundable, log in, open "Get help", send a message, choose "Claude Refund Request" and follow the prompts to check eligibility. Customers in the European Economic Area and the UK can get a refund prorated to usage within 14 days of purchase, and subscriptions bought through the App Store on iOS can only be refunded by Apple. As of 2026-11-12, refunds follow whoever took the payment: a subscription bought on Anthropic's website is handled under the Consumer Terms, while one bought through a distributor such as the App Store follows that distributor's payment and refund terms; see /claude-subscription-ban-refund for the step-by-step comparison.

Your account data: export it first

According to the Help Center (checked 2026-11-09), Free, Pro, Max and Ultra accounts banned for a Usage Policy violation can still log in at claude.ai to export data or delete the account, without contacting Anthropic, although what you can export may be restricted depending on the violation. The export lives under Settings, then Privacy, then Export data; the download link arrives by email and expires 24 hours after delivery.

The Consumer Terms also allow Anthropic, at its option, to delete data associated with the account once access is terminated, so export early. Your Claude Code project code lives in your own repositories and does not disappear with the account; what needs saving is mainly your claude.ai conversations and project material.

Subscription login vs API credentials: two ways to connect

Claude Code connects in one of two ways: a claude.ai subscription login, or an API credential pointed at a gateway. The official Claude Code docs explain that while a credential variable is active, requests carry that credential instead of the claude.ai subscription login. Using usage-billed API credentials for team, automation and production work is the idea behind the comparison table above. A banned account goes through the official appeal described above; Anthropic's Usage Policy explicitly prohibits circumventing a ban through a different account, new or existing.

On QCode: QCode.cc is an independent API access service, not affiliated with or endorsed by Anthropic. To connect, set ANTHROPIC_BASE_URL to https://api.qcode.cc/api and put your cr_ key in ANTHROPIC_AUTH_TOKEN; the same key calls claude-sonnet-5, claude-opus-5-5, claude-sonnet-5-5 and more, with the full list on /models. Billed per token; per-model rates are listed on /models. As of 2026-11-18, QCode also supports calling claude-fable-5-2, claude-fable-5-1, claude-haiku-5-5, gpt-6.2-sol, gpt-6.2-astra and qwen4-max (use https://api.qcode.cc/openai for Codex CLI or https://api.qcode.cc/openai/v1 for the OpenAI SDK, Cline and OpenCode).

  • Subscription login: suited to personal interactive use, and unusable once the account is disabled.
  • API credential: Claude Code connects with ANTHROPIC_BASE_URL plus ANTHROPIC_AUTH_TOKEN and is billed per token.
  • Both side by side: the credential variable takes precedence over a saved claude.ai login, and unsetting it returns Claude Code to the subscription.

Questions after a ban

How long does a ban appeal take?

Anthropic publishes no fixed timeframe. As of 2026-11-09 (the same on 2026-10-15), the Help Center's appeal page still says response times are currently longer than normal and that appeals will be answered as soon as possible. While you wait, you can log in to claude.ai and export your data.

Can I get a refund for Claude Pro or Max after a ban?

Not if access was terminated for a violation of the terms, according to the Consumer Terms; if Anthropic ends a website-purchased subscription for another reason, it refunds the remaining time pro rata. EEA and UK customers also have a prorated refund within 14 days of purchase, and App Store subscriptions are refunded by Apple. For a payment the terms define as refundable, send a message under "Get help" in Claude and choose "Claude Refund Request".

Can I still export my chats after a ban?

Usually yes. The Help Center says all banned users can delete their own account and self-serve exports are available to all banned users, though the data you can export may be restricted depending on the violation; start the export under Settings, then Privacy, and use the link within 24 hours.

Can Claude Code only be used with a subscription login?

No. The official docs support connecting to a gateway with ANTHROPIC_BASE_URL and ANTHROPIC_AUTH_TOKEN, and the credential variable takes precedence over a saved claude.ai login. This is a standard, documented configuration option; a banned account goes through the official appeal, and Anthropic's Usage Policy applies whichever way you connect. A QCode cr_ key connects this way; billed per token, with per-model rates on /models.

How does a QCode key relate to my Claude subscription?

They are independent. The QCode docs state that an Anthropic subscription cannot be used on QCode and is never drawn down by it, and QCode usage does not touch your official subscription's allowance; the two are parallel systems.

How do I pay for and activate QCode from mainland China?

Top up a CNY balance with Alipay or WeChat Pay, or use Visa / Mastercard, Apple Pay, Google Pay or crypto across many coins and networks. You top up first and then buy a plan, and your key starting with cr_ is created as soon as the plan is bought.

Per-token API access

One key for Claude, GPT and leading Chinese models: API access billed per token, with usage you can watch.

View plans

Try first, then decide

Not sure which tier? Start with Starter ($9.00/mo) and upgrade when you're happy — the unused value of the old plan goes back to your balance.