Acceptable Use Policy

You may not use the Services to generate, store, distribute, or solicit any content that:

• depicts or sexualizes minors, or facilitates child sexual abuse material (CSAM) in any form;
• provides operational instructions for weapons capable of mass casualties (chemical, biological, radiological, nuclear, or high-yield explosives);
• facilitates planning, executing, or evading prosecution for violent crimes against specific persons or groups;
• generates non-consensual sexual imagery, deepfakes used to harass or defame, or imagery designed to deceive about the identity of a real person;
• is intended to defraud, deceive, or financially harm specific individuals (e.g. phishing kits, romance-scam scripts);
• promotes self-harm, eating disorders, or suicide in a manner contrary to widely-accepted clinical guidance;
• infringes third-party intellectual property, trade secrets, or privacy rights;
• violates applicable export control or sanctions law.

You may not:

• share, resell, or redistribute API keys without our written permission;
• attempt to deanonymize, identify, or directly contact our underlying inference providers, or to bypass MWS to reach them;
• attempt to extract our model weights, prompts, configuration, or internal routing logic;
• use the Services to train competitive products without our written permission;
• generate traffic patterns designed to evade rate limits or exhaust shared resources for other customers;
• exceed 100 requests per second from a single account without prior written approval (per-key rate limits start lower; see the docs).

The Services are designed for software development, prototyping, and general-purpose AI tasks. They are not designed or warranted for use in:

• medical diagnosis, prescribing, or treatment decisions affecting patient care;
• legal advice provided to clients without a qualified attorney's review;
• financial advice provided to consumers without a qualified advisor's review;
• fully-automated decisions affecting employment, credit, insurance, housing, or government benefits;
• control of critical infrastructure (power, water, transport, defense systems);
• any application where a model error could cause significant physical, financial, or psychological harm without meaningful human review.

If you intend to use the Services for anything in this category anyway, you do so at your own risk and you must implement qualified human review, audit logging, and an off-switch.

Model output is statistical, not factual. It can be wrong, biased, fabricated (“hallucinated”), out of date, or unsafe. You are responsible for reviewing output before relying on it, publishing it, or acting on it.

We do not warrant that any output is accurate, complete, fit for any particular purpose, or non-infringing. We do not indemnify you against claims arising from your use of model output.

Report abuse, security issues, or AUP violations to support@vellora.ai. We investigate every report and respond within five business days. Urgent safety issues (e.g. exposed key, suspected account takeover) are triaged within 24 hours.

We reserve the right, but do not assume the obligation, to investigate suspected AUP violations. Enforcement actions may include:

• warning the account holder;
• throttling or suspending API access;
• revoking specific API keys;
• terminating the account;
• forfeiting unused credits (in cases of severe or repeated violation);
• reporting illegal activity to law enforcement, including preserved logs as required by law.

We will provide reasonable notice for non-urgent enforcement actions where we can do so without aiding the violation.

We may update this AUP from time to time. Material changes will be announced via email to account holders at least 14 days before they take effect. Continued use of the Services after the effective date constitutes acceptance.