> ## Documentation Index
> Fetch the complete documentation index at: https://docs.globalai.vip/llms.txt
> Use this file to discover all available pages before exploring further.

# Global AI Usage Policy: Regional Limits and Compliance

> Global AI regional restrictions, prohibited uses, and compliance rules. Read before use; violations result in account suspension.

Global AI provides AI model aggregation for compliant use cases. Read the following regional restrictions and usage rules before using the platform. Registering or making calls means you agree to this policy. Accounts that violate the rules will be suspended.

## Regional restrictions

<Warning>
  Global AI does not provide service to users in the countries and regions listed below. Do not use proxies, VPNs, or other means to bypass these restrictions. Any account suspension that results is your own responsibility.
</Warning>

Upstream providers such as OpenAI and Anthropic (Claude) do not offer service in certain countries and regions. As an aggregation service, Global AI likewise does not open access to users in those regions, including but not limited to:

* Iran
* North Korea
* Cuba
* Syria
* China (including Hong Kong, Macau, and Taiwan)
* Crimea
* Russia
* Other countries and regions sanctioned or restricted by the US OFAC or the relevant upstream providers

If registration or calls from these regions are detected, the platform reserves the right to:

* Refuse service
* Suspend or close the account
* Halt top-ups and refunds
* Pursue legal liability where applicable

## Prohibited uses

The following uses are prohibited on Global AI. When a violation is found, the platform immediately disables the affected tokens and account, refuses refunds, and cooperates with regulators when required.

<AccordionGroup>
  <Accordion title="Illegal activity" icon="ban">
    Any use that violates the laws and regulations of the user's country, including but not limited to fraud, gambling, drug trafficking, money laundering, terrorism, and smuggling.
  </Accordion>

  <Accordion title="Violence and harm" icon="triangle-exclamation">
    Generating or distributing content involving serious violence, gore, self-harm, harm to others, or content that endangers public safety. Planning, inciting, or assisting any form of real-world harm.
  </Accordion>

  <Accordion title="Sexual content and child protection" icon="user-shield">
    Generating or distributing pornographic, softcore, or sexually exploitative content. **Any inappropriate content involving minors is strictly prohibited**, regardless of form or purpose.
  </Accordion>

  <Accordion title="Political sensitivity" icon="flag">
    Generating or distributing content involving national sovereignty, territorial integrity, or politically sensitive topics. Content that incites subversion, damages national unity, endangers national security, or violates local laws.
  </Accordion>

  <Accordion title="Hate and discrimination" icon="hand">
    Hate speech, personal attacks, harassment, or discriminatory content based on race, ethnicity, nationality, religion, gender, sexual orientation, disability, age, or similar attributes.
  </Accordion>

  <Accordion title="Privacy and fraud" icon="user-secret">
    Scraping, selling, or leaking others' private information. Identity forgery, impersonation, phishing, or social-engineering fraud. Forging IDs, transcripts, official documents, and the like.
  </Accordion>

  <Accordion title="Malicious use and abuse" icon="bug">
    Writing malicious code, viruses, ransomware, or penetration tools. Conducting DDoS attacks, scraper abuse, reverse engineering, or risk-control bypasses against Global AI or any third party. Bulk registration, quota farming, sharing or reselling tokens, and other actions that disrupt normal platform operations.
  </Accordion>

  <Accordion title="High-risk decisions" icon="stethoscope">
    Directly applying model output in high-risk contexts (medical diagnosis, legal rulings, financial investment, life safety) without review by a qualified professional. AI output is for reference only and cannot replace professional judgment.
  </Accordion>
</AccordionGroup>

## Abuse controls

Beyond prohibited uses, Global AI also strictly controls **resource abuse** that harms platform stability, fairness, or commercial order. The following scenarios are treated as abuse even when the underlying content is lawful.

### Common abuse scenarios

<AccordionGroup>
  <Accordion title="Account and token abuse" icon="user-slash">
    * Bulk-registering accounts with false information, or farming welcome bonuses or referral rewards across multiple accounts
    * Sharing, renting, reselling, or publicly leaking accounts or tokens
    * Using the same phone, email, device, or IP range to bypass quota or bans
    * Illegally reselling Global AI accounts or tokens on third-party platforms
  </Accordion>

  <Accordion title="Quota and billing abuse" icon="coins">
    * Cheating to obtain trial quota, bonus quota, or promotional rewards
    * Malicious refunds, chargebacks, stolen-card payments, or top-ups via illegal payment channels
    * Exploiting system bugs, timing gaps, or concurrency races to arbitrage
    * Fraudulent redemption, forged codes, or reuse of single-use credentials
  </Accordion>

  <Accordion title="Request and traffic abuse" icon="bolt">
    * Sustained ultra-high concurrency calls, or scripted load tests and scans with no meaningful purpose
    * Loop, infinite-loop, or unbounded Agent and automation calls
    * Large-scale content scraping, data distillation, or training competing models
    * Repackaging the Global AI API and reselling it as a third-party API service without authorization
    * Deliberately padding context or accumulating useless history messages to inflate token billing disputes
  </Accordion>

  <Accordion title="Platform attacks and evasion" icon="shield-halved">
    * DDoS, CC attacks, scraping, scanning, or penetration testing against Global AI or its upstream services
    * Reverse engineering, decompiling, or bypassing rate limits, authentication, risk control, or billing
    * Bypassing regional restrictions or bans through proxies, VPNs, or forged IPs
    * Malicious reports, false accusations, forged evidence used to harass support or disrupt operations
  </Accordion>

  <Accordion title="Content-layer abuse" icon="robot">
    * Using prompt injection or jailbreak prompts to bypass model safety policies
    * Coercing or inducing models to produce content listed in this policy's [Prohibited uses](#prohibited-uses)
    * Using output to generate spam email, spam comments, SEO cheating, astroturfing, or fake public opinion
    * Creating deepfakes, face swaps, or forged audio and video for misleading purposes
  </Accordion>
</AccordionGroup>

### Tiered enforcement

The platform combines log analysis, behavioral modeling, and upstream feedback to detect abuse, then applies the following measures based on severity:

| Level        | Triggers                                                              | Enforcement                                                                          |
| ------------ | --------------------------------------------------------------------- | ------------------------------------------------------------------------------------ |
| Notice       | Short bursts of ultra-high concurrency, unusual call patterns         | System warning and temporary throttling; auto-lifted once behavior returns to normal |
| Restrict     | Repeated rate-limit hits, suspected arbitrage, suspected sock puppets | Lower rate limit, restrict model access, disable some features                       |
| Freeze       | Clear abuse, reselling, bypassing risk controls                       | Freeze the token and account balance, halt top-ups and withdrawals                   |
| Ban          | Severe abuse, attacks on the platform, suspected illegal activity     | Permanent ban on the account and any linked accounts, with no refund                 |
| Legal action | Actual damages or suspected criminal activity                         | Pursue civil damages and refer the case to law enforcement                           |

<Note>
  Legitimate high-concurrency needs are not abuse. If you have production requirements for high quota or high concurrency, [contact us](https://app.mintlify.com/globalai/globalai/~/a30d2b88-b5dc-4d57-a6eb-9d9ed5a3a97b) in advance to raise your limits and avoid false positives from the risk-control system.
</Note>

## Violation handling

The platform handles violations as follows:

| Severity        | Response                                                                     |
| --------------- | ---------------------------------------------------------------------------- |
| Minor           | Warning and remediation request. Restrict some models or quota               |
| Standard        | Disable affected tokens, freeze the account balance, refuse refunds          |
| Severe          | Permanent account ban. Reserve the right to pursue legal action              |
| Suspected crime | Refer to law enforcement, cooperate with the investigation, provide evidence |

## Have a question?

If you are unsure whether a specific use complies with this policy, [contact us](https://app.mintlify.com/globalai/globalai/~/a30d2b88-b5dc-4d57-a6eb-9d9ed5a3a97b) before you begin. We can help you assess the compliance risk and avoid account losses due to misuse.


## Related topics

- [Global AI FAQ: Accounts, Tokens, Calls, and Billing](/en/faq.md)
- [Global AI Token Management: Create, Limit, and Secure](/en/usage/tokens.md)
- [Usage Logs & Dashboard: Inspect Calls and Trends](/en/usage/logs.md)
- [Call Global AI: OpenAI, Claude, and Gemini Formats](/en/usage/api.md)
- [Cline & Roo Code with Global AI: Coding Agent Guide](/en/clients/coding-agents.md)