This Usage Policy applies to the AIgateway and AI Gateway services operated by Roushan, Inc. (Dover, Delaware, USA) and is incorporated by reference into our Terms of Service. AIgateway relays requests to upstream model providers. You are responsible for the prompts you send and the outputs you use. This policy sets the floor — individual providers may have stricter rules that we will enforce on their behalf.
1. Prohibited uses
You must not use the Service to:
- Produce or distribute Child Sexual Abuse Material (CSAM) or content that sexualizes, grooms, or exploits minors. Any detected CSAM is reported to NCMEC as required by U.S. federal law (18 U.S.C. § 2258A).
- Generate material intended to harass, threaten, defame, or impersonate a specific real person without consent.
- Design, synthesize, plan, or operationalize weapons capable of mass casualties (biological, chemical, nuclear, radiological), or promote, provide logistical support for, or glorify terrorism.
- Plan or execute attacks on, intrusions into, or denial-of-service against computer systems you do not own or have explicit permission to test.
- Generate, distribute, or operationalize malware, ransomware, phishing kits, or credential-harvesting infrastructure.
- Commit fraud, including generating fake identity documents, forged signatures, synthetic audio for scams, or content intended to mislead in financial, securities, or benefits contexts.
- Run automated decisions that materially affect a person's legal rights or access to essential services (loans, hiring, housing, parole, insurance, healthcare) without meaningful human review and required disclosures.
- Build political-persuasion bots that misrepresent themselves as human, generate deceptive synthetic media (“deepfakes”) of real people without their consent, or interfere in elections.
- Generate sexual content involving real people without their consent, non-consensual intimate imagery, or content that violates image-based sexual abuse laws in the user's jurisdiction.
- Scrape content or credentials in violation of a website's terms of service, robots directives, or applicable law (CFAA, DMCA, national equivalents).
- Infringe the intellectual-property or privacy rights of others, or knowingly violate export-control or sanctions laws (see the Export controls section of our Terms).
- Resell raw upstream access without adding substantial value and without disclosing your relationship with AIgateway.
- Circumvent usage limits, rate limits, guardrails, spend caps, or abuse-prevention measures.
2. Use carefully
- High-stakes advice. Medical, legal, mental-health, and financial outputs must be labelled as AI-generated and reviewed by a qualified human before a user acts on them. Do not substitute the Service for a licensed professional.
- Code execution. Sandbox model-generated code. Treat it as untrusted input.
- User-generated prompts. Rate-limit them, tag them with
x-aig-tag, and enforce a spend cap per sub-account. If you pass end-user prompts straight through without filtering, you accept the full upstream-provider policy for that content. - Synthetic media. Where your jurisdiction requires it, label AI-generated images, audio, or video as such.
- Evaluations + benchmarks. Publishing comparative benchmarks of upstream providers is fine; representing upstream output as your own frontier model is not.
3. Children's safety
If you build a product accessible to users under 13 (or under the age of digital consent in their jurisdiction), you must comply with COPPA, GDPR-K, India DPDP Act children's provisions, and equivalent local law, and you must not send personal information about those users through the Service except as strictly necessary for the feature.
4. Reporting abuse
If you see output that violates this policy, suspect another customer is abusing the platform, or encounter unlawful content, email abuse@aigateway.sh. We triage within 24 hours and act within 72. Urgent safety reports (imminent harm, CSAM, targeted threats) may be escalated to trust@aigateway.sh or law-enforcement as required.
5. Enforcement
We may suspend or terminate access without notice for active abuse. For gray-area cases we will contact you first. Suspensions do not entitle you to a refund of usage already incurred. Repeat or severe violations may result in permanent termination and IP/device bans. We will cooperate with law-enforcement under lawful process.
6. Upstream provider policies
Some providers ban use cases we allow (for example, adult content on specific models) and some allow use cases we do not. When you select a model, you accept that provider's terms for the duration of the request. The models page flags provider-specific restrictions. A provider-banned prompt may be blocked, logged, or reported to the provider even if it does not violate this policy.
7. Changes to this policy
We update this policy as new risks, regulations, and upstream-provider rules emerge. Material changes are announced at least 30 days in advance via the changelog and email, except where an immediate change is necessary to protect users or comply with law.
8. Contact
Abuse reports: abuse@aigateway.sh
Policy questions: legal@aigateway.sh
Mailing address: Roushan, Inc., 1111b S Governors Ave, STE 55131, Dover, DE 19904, USA.