Acceptable Use Policy

Scope

This Acceptable Use Policy (“AUP”) applies to every use of the Modelith website, dashboard, API, webhooks, and any other service we operate. It supplements the Terms of Service. Where a single act could be a violation of both this AUP and the Terms, both apply.

You are responsible for the acts of anyone using your account or your API keys, whether or not you authorized the use. If you operate a service that sends traffic to Modelith on behalf of end users, you must enforce this AUP against your end users and remain liable for their violations.

Prohibited content

You may not use the service to generate, transmit, store, or process any of the following categories of content:

• Content that is illegal under the laws of the jurisdiction in which you reside, the jurisdiction in which the content is produced, or the jurisdiction in which it is received
• Child sexual abuse material (CSAM) in any form, including content that depicts, describes, or encourages the sexual exploitation of minors
• Content that facilitates the planning or execution of violence against a person, a group, or a protected category, including terrorist content and material that praises or depicts genocide
• Non-consensual intimate imagery, doxxing material, or content designed to harass, stalk, or threaten a specific individual
• Material that infringes the intellectual property rights of a third party, including copyrighted text, images, and source code reproduced beyond fair use
• Content designed to deceive a person about its origin, its authenticity, or its purpose (including deepfakes and undisclosed synthetic media used to manipulate a person)
• Spam, including unsolicited bulk messages, comment spam, and SEO link schemes
• Content that assists in the design, production, or use of weapons of mass destruction, or that provides operational guidance for violent non-state actors

Prohibited conduct

You may not, in connection with the service, engage in any of the following:

• Prompt injection against the service.Sending instructions to a model that is designed to subvert, manipulate, or extract the system prompt, the routing rules, the wallet state, or another customer’s metadata through shared infrastructure.
• Bypassing rate limits or billing. Creating multiple accounts, rotating API keys, or otherwise distributing traffic to evade per-key, per-user, or per-IP rate limits or to consume platform routing without paying for it.
• Reverse engineering the governor. Probing, scanning, fuzzing, or otherwise attempting to extract the routing engine, the complexity scorer, the pricing model, or the wallet logic at scale. Routine use of the public API to learn its behaviour is not reverse engineering.
• Credential abuse. Sharing API keys publicly (in source repositories, in client-side code, in support tickets), or attempting to authenticate as another customer.
• Service degradation. Denial-of-service attacks, traffic flooding, or automated workloads that materially degrade the platform for other customers.
• Reselling without authorization. Reselling, sublicensing, or white-labeling access to the Modelith routing API as a competing routing product without a written partnership agreement.
• Harm to others. Using outputs from the service to facilitate any of the harms listed in the Prohibited Content section above.

Platform integrity

The following additional rules protect the platform itself. They are not negotiable:

• You may not probe, scan, or test the security of modelith.cloud except through the vulnerability disclosure channel described in the Security page.
• You may not use the service to train, fine-tune, or improve a competing routing product or a competing cost-governance plane.
• You may not place automated workloads on the free tier that are designed to extract platform routing at a discount; the free tier is a developer evaluation tool, not a production platform.
• You may not misrepresent your identity, your affiliation, or the purpose of your traffic to Modelith staff, to upstream providers, or to other customers.

Enforcement ladder

When we identify a violation, we apply the following ladder. The severity of the violation, the customer’s history, and the risk to the platform determine the rung we start on.

1. Warning. Email to the account holder describing the issue and the required remediation. Suitable for low-severity, first-time, easily reversible issues.
2. Throttling. Reduced rate limits, partial routing, or queueing of requests. Applied when a customer is generating excessive load or operating outside the declared use case.
3. Suspension. Full suspension of the account or specific API keys for a defined period. Applied for repeated violations or for issues that require customer action to resolve.
4. Termination. Permanent closure of the account and revocation of all API keys. Applied for severe, repeated, or unremediated violations, and for violations involving CSAM, violence, or other categories listed in the Prohibited Content section.
5. Reporting. For violations that involve illegal activity or that require reporting under applicable law, we will report the violation to the appropriate authorities and to the affected upstream providers. We will also preserve evidence as required.

We may skip rungs of the ladder when the violation is severe, when the risk to the platform or to other customers is acute, or when preservation of evidence is required by law.

Reporting violations

If you believe a customer of Modelith is violating this AUP, or if you have received Modelith-routed output that you believe is prohibited, email founders@modelith.cloud with the subject line “AUP report”. Include as much context as you can: the API key prefix, the request ID, the timestamp, the model involved, and a description of the issue. We acknowledge AUP reports within 1 business day and aim to take a first action within 5 business days.

For reports of child sexual abuse material, do not forward the material to us. Report directly to the National Center for Missing and Exploited Children (NCMEC) CyberTipline in the United States, or to the equivalent authority in your jurisdiction. We will cooperate with law enforcement on preservation and disclosure requests.

Compliance with law

You are responsible for understanding and complying with the laws that apply to your use of the service. This includes, where applicable, export control regulations (including the U.S. Export Administration Regulations and the EU Dual-Use Regulation), data protection law (including GDPR and CCPA), sector-specific regulation (including HIPAA where applicable), and the terms and acceptable-use policies of every upstream provider whose models you route through Modelith.

Nothing in this AUP is legal advice, and nothing in this AUP limits your obligations under any law that applies to you. If a provision of this AUP conflicts with a law that applies to you, the law controls, but only with respect to the conflicting part and only for the customer and the jurisdiction at issue.