Terms of Service

2. Geographic Restrictions

IMPORTANT: SIToolBase is not available for use in Mainland China.

Our Services are not intended for distribution to, or use by, any person or entity in Mainland China. By accessing or using our Services, you represent and warrant that:

• You are not located in Mainland China
• You are not a resident of Mainland China
• You will not access or use our Services from within Mainland China
• You are not subject to the jurisdiction of Mainland China

If you are located in or a resident of Mainland China, you are expressly prohibited from accessing or using our Services. We reserve the right to block access from IP addresses originating from Mainland China and to terminate accounts of users found to be in violation of this restriction.

4. Prohibited Content, Reporting and Content Moderation

IMPORTANT: Violation of these rules may result in content removal, credit forfeiture, account suspension or permanent termination.

This section summarizes the content rules that apply to every user. The full content policy — including our moderation system, severity levels, response timelines and appeals — is set out in our Acceptable Use Policy (AUP), which forms part of these Terms.

4.1 The six prohibited categories

You agree NOT to use our Services to create, upload, process or distribute any of the following six categories of content:

1. Sexual / NSFW content — pornography, nudity, obscene or sexually explicit material, including AI-generated adult content.
2. Violent / gory content — graphic violence, gore, self-harm or content that glorifies harm to people or animals.
3. Hate speech — content attacking, threatening, insulting or dehumanising people or groups on the basis of race, ethnicity, nationality, religion, gender, gender identity, sexual orientation, disability or any other protected characteristic.
4. Child sexual abuse material (CSAM) — any content that sexualises, endangers or exploits minors, or that is otherwise unsafe for minors.
5. Deepfakes / impersonation of others — synthetic media depicting real, identifiable people without their explicit consent, deceptive deepfakes, and impersonation of any person, brand or institution.
6. Copyright or trademark infringement — content that violates intellectual property, image, personality or privacy rights.

4.2 Zero tolerance — immediate termination

The following are prohibited without exception. Content in this category is removed immediately, the responsible account is permanently terminated without refund, and we may report the matter to law enforcement:

• Child sexual abuse material (CSAM), or any sexualised content involving minors
• Content promoting, instructing or facilitating terrorism, violent extremism or mass violence
• Instructions for manufacturing weapons of mass destruction
• Content inciting genocide, racial or ethnic hatred
• Non-consensual sexual deepfakes depicting real, identifiable people

4.3 Detailed prohibitions

Without limiting the categories above, you also agree NOT to create, upload, process or distribute:

• Illegal Content: any content that violates local, state, national or international laws
• Non-Consensual Content: AI-generated or edited content involving individuals without their explicit consent
• Harmful Content: content that promotes violence, hatred, discrimination or harassment
• Defamatory Content: content that damages the reputation of any person or organization
• Deceptive Content: deepfakes intended to mislead, defraud or spread misinformation, or AI output presented as human-made in order to deceive
• Political Content: AI-generated or edited content involving political figures, government officials or candidates for public office, especially content that could influence elections or public opinion
• Religious Content: content that mocks, disparages or misrepresents religious figures, symbols, practices or beliefs in a manner that could cause offense or controversy
• Racial Discrimination Content: any content that promotes, encourages or depicts racial discrimination, racial stereotypes or racial hatred
• Ethnic Discrimination Content: content that targets, discriminates against or perpetuates harmful stereotypes about specific ethnic groups
• Controversial Public Figures: content involving celebrities, public figures or influencers in contexts that could cause controversy, damage reputations or misrepresent their views or actions
• Government/Military Content: content involving military personnel, government officials or state symbols in inappropriate or misleading contexts
• Platform Abuse: bulk or spam generation, fraudulent use, jailbreak prompts, or any attempt to bypass, disable or evade our content moderation systems

4.4 How to report violating content

If you find content or behaviour that violates this section, please report it to us:

• Report email: [email protected] — subject line "Content report"
• Urgent violations (child safety, terrorism, imminent harm or non-consensual intimate content): same address with the subject line "URGENT — Content Safety"
• What to include: the link or task ID of the content (or the account involved), a short description of why it violates this section, and your contact details if you would like a reply
• Confidentiality: the identity of a reporter is kept strictly confidential and is not disclosed to the reported user
• Rights holders: if you are the person depicted, or you own the rights to the material, say so in your email and attach supporting evidence — such reports are prioritised

4.5 Content moderation mechanism

We operate a layered moderation system combining automated filtering, human review and periodic sampling:

• Input moderation (before generation): prompts, style descriptions and other free text are screened by an OpenAI-compatible moderation model (omni-moderation-latest) before a task is created. Zero-tolerance categories, including sexual/minors and self-harm/instructions, are blocked unconditionally; other categories are blocked above a confidence threshold. Blocked requests are rejected before any generation and no credits are deducted. A maintained blocklist of prohibited terms (sexual/NSFW, CSAM-related, copyright/brand/trademark, violence, hate, self-harm, deepfake/celebrity/political) is also applied to free-text input and inside our prompt-building layer.
• Output moderation (after generation): generation is performed by the third-party foundation models listed in section 5; every result passes through those models' safety systems before it is returned, and outputs rejected by those systems are not delivered to users.
• Human review: reported content is reviewed by our trust & safety team, and content flagged by automated filters may be queued for human confirmation.
• Sampling and audit: recent tasks and moderation logs are reviewed on a sampled basis, and moderation thresholds and blocklists are updated regularly.
• Logging: moderation decisions, including category scores, block reasons and affected task IDs, are retained for abuse investigation and to respond to law-enforcement or payment-partner requests.

4.6 Response and resolution timelines

If a case needs more time we will notify the reporter and explain why. Enforcement actions range from a written warning and content removal, through feature limitations and credit forfeiture, to temporary suspension, permanent ban and referral to law enforcement. Appeals may be submitted to [email protected] (subject "Appeal") within 30 days and are reviewed by a reviewer who was not involved in the original decision. Severity levels, enforcement measures and the appeals process are described in full in the Acceptable Use Policy.