Acceptable Use Policy

We build KirinAI to help compliance-aware teams ship safe assistants. This policy translates that intent into unambiguous rules. It applies to you, your authorised users, and any end-user traffic routed through agents you deploy on our platform. Violations may lead to warnings, suspension, or termination under Section 7 of our Terms of Service.

You may not use KirinAI to produce, store, or transmit content that:

• Depicts or sexualises minors.
• Incites or facilitates violence against individuals or protected groups.
• Provides operational guidance for the creation of weapons capable of mass harm (biological, chemical, nuclear, radiological, or cyber-kinetic weapons).
• Is used to harass, stalk, dox, or threaten any individual.
• Violates intellectual property or privacy rights of a third party.

You may not:

• Probe, scan, or penetrate the service without prior written authorisation.
• Bypass rate limits, quotas, or authentication controls.
• Attempt to extract model weights, training data, or other IP by prompt injection or otherwise.
• Upload malware, worms, or trojans.

You may not:

• Generate or distribute political disinformation, election manipulation, or impersonation content.
• Misrepresent agent output as human-authored when the context requires disclosure (e.g. regulated advice, educational credentials).
• Operate spam farms, content-mill SEO operations, or credential-stuffing pipelines.

KirinAI outputs are probabilistic and not a substitute for licensed professionals. You may not present agent responses as legal, medical, tax, or financial advice without adding clear disclaimers, human review, and jurisdiction-appropriate compliance controls.

Only upload third-party personal data that you have a lawful basis to process. If you submit customer data on behalf of your own clients, you are responsible for obtaining appropriate notice and consent from those clients.

Credits purchased under your subscription are for operating agents that serve legitimate business use cases. Reselling raw model access, running proxy services, or automating large-scale scraping is not a permitted use of credits, even if they are within your quota.

If you believe an agent on KirinAI is violating this policy, please report it to abuse@kirinai.com. Include URLs, timestamps, and any relevant conversation IDs. We respond to reports within two business days.

We investigate violations case by case. Minor or first-time infractions usually result in a warning. Severe, repeated, or clearly malicious violations lead to suspension without warning. Where required, we cooperate with law enforcement.