# Acceptable Use Policy Last updated: 2026-06-23 --- This Acceptable Use Policy (AUP) is an integral part of the KeepRouter [Terms of Service](/terms). By using the Service you agree to comply with this AUP. Violations may result in content blocking, throttling, account suspension or termination, and in serious cases we will report to the relevant authorities as required by law. ## 1. Scope This AUP applies to all of your access to and use of every upstream model (OpenAI, Anthropic/Claude, Google Gemini, as well as Qwen, Kimi, GLM, DeepSeek, MiniMax, etc.) through the Service, and to all authorized users and end users under your account. ## 2. Compliance With and Flow-Down of Upstream Policies 2.1 **Core obligation:** You must comply with and flow down the usage policies of every upstream provider you call (including but not limited to the OpenAI Usage Policies, the Anthropic Usage Policy, the Google Generative AI Prohibited Use Policy, and the corresponding policies of domestic model providers). 2.2 If you integrate the Service into your own product for end users, you must bind your end users to standards **no less strict** than this AUP and each upstream policy. Where multiple policies conflict, the stricter applies. 2.3 Upstream policies may be updated at any time, and different models are subject to different policies; it is your responsibility to review the latest version before calling. ## 3. Prohibited Uses You may not use the Service, or assist or permit others to use it, to: - **Unlawful content and activity:** any content or conduct that violates applicable laws and regulations. - **Child sexual abuse material (CSAM) — zero tolerance:** generating, distributing, soliciting, obtaining or in any way involving CSAM, the sexualization of minors, or the grooming of minors is strictly prohibited, whether or not AI-generated. We have zero tolerance for this and will immediately terminate the account, preserve relevant evidence, report to the relevant authorities (such as NCMEC or the local counterpart) and cooperate with law enforcement as required by law. - **Harm to people and public safety:** promoting suicide, self-harm or eating disorders; planning or carrying out violence, terrorism or violent extremism; harassment, threats, bullying or hate speech. - **Weapons and CBRN:** developing, designing, acquiring or using weapons, explosives, or chemical, biological, radiological or nuclear (CBRN) materials. - **Critical infrastructure:** disrupting or gaining unauthorized access to critical infrastructure or safety systems such as power, water, transport, finance, healthcare or elections. - **Cyber-attacks and malicious code:** creating or distributing malware, ransomware or botnets; launching denial-of-service attacks; unauthorized probing, scanning or intrusion of the Service, upstream systems or any third-party system. - **Abuse, fraud and malicious behavior:** fraud, phishing, spam, fake reviews, impersonation, social engineering, pyramid/Ponzi schemes, fabricated or manipulated media (deepfakes) used to deceive. - **Privacy and surveillance:** unauthorized collection or disclosure of personal information; misuse of facial/biometric recognition; unlawful mass surveillance. - **Elections and political manipulation:** mass-generating political lobbying content that conceals its AI origin, voter-suppression or disinformation campaigns. - **Circumventing technical and billing limits:** attempting to bypass or break rate limits, quotas, or billing/credit deduction; unauthorized red teaming. - **Unauthorized resale:** reselling, subletting or redistributing the Service's API access without our written authorization. - **Infringement:** infringing others' intellectual property, privacy, likeness or other rights. - **High-risk and restricted uses:** providing personalized medical, legal, financial or other advice requiring professional qualifications without independent review by a licensed professional (human-in-the-loop), or using Outputs for automated decisions with significant legal or similar effects on individuals (such as credit, employment, housing, education, justice/law enforcement); and other uses generally prohibited by upstreams. - **Age and region:** for use only by those aged 18 and older; must not be used within mainland China, and must not circumvent upstream geographic / export-control restrictions. ## 4. Your Responsibilities & End-User Deployment 4.1 You are responsible for all requests and content originating from your account and keys, must ensure you have all rights needed to submit Inputs, and must verify the use of Outputs as necessary. 4.2 **AI disclosure:** If you deploy a chatbot or similar interactive AI application to consumers, you must clearly inform end users that they are interacting with AI (not a human), to the extent required by applicable law (such as the EU AI Act) and upstream policies, and provide appropriate human review and disclaimers where high-risk domains are involved. 4.3 You must provide your end users with acceptable-use terms consistent with this AUP and each upstream policy, and take reasonable steps to stop abuse when you discover it. ## 5. Enforcement 5.1 We may use automated and manual means to monitor abuse. On discovering a violation of this AUP or any upstream policy, we may, without prior notice, block requests, throttle, suspend or permanently terminate the account, claw back the credit involved in the violation, and report to upstream providers and the relevant authorities where necessary. 5.2 For termination due to a violation, remaining credit may be non-refundable. 5.3 We may take enforcement measures at the request of upstream providers or as required by law. ## 6. Reporting If you discover any violation of this AUP (especially involving CSAM or an urgent safety risk), report it immediately to support@keeprouter.com; reports involving CSAM or urgent threats to personal safety will be handled with priority and urgency. ## 7. Changes We may update this AUP from time to time; continued use after changes take effect constitutes acceptance. ## 8. Contact support@keeprouter.com