Acceptable Use Policy

Effective date: 2026-10-07 · Last updated: 2026-10-07

This Acceptable Use Policy ("AUP") sets out the rules for using the Zyverna website, application, agent features and API (the "Service") provided by Zyverna, Inc.. It forms part of our Terms of Service. Capitalized terms have the meanings given there.

We want Zyverna to be a safe, reliable tool for building software. If you are unsure whether a use is permitted, ask us at [email protected] before proceeding.

1. Prohibited uses

You may not use the Service, or allow others to use it, to:

1.1 Break the law or harm others

• violate any applicable law, regulation or court order, including export control and sanctions laws;
• infringe or misappropriate intellectual property, privacy, publicity or other rights;
• generate or distribute defamatory, fraudulent, deceptive or misleading content, including impersonating a person or organization;
• harass, threaten, bully, intimidate or incite violence against any person or group;
• generate content that sexually exploits or endangers children in any form (we report such content to the National Center for Missing & Exploited Children and relevant authorities);
• promote self-harm, terrorism, violent extremism or human trafficking.

1.2 Create or deploy harmful software

• develop, test or distribute malware, ransomware, spyware, exploits, or tools whose primary purpose is unauthorized access, denial of service, credential theft or evasion of security controls;
• conduct unauthorized penetration testing, scanning or attacks against systems you do not own or have explicit permission to test;
• create automated systems to scrape, spam, flood or abuse third-party services in violation of their terms.

1.3 Deceive or manipulate

• produce disinformation, coordinated inauthentic behavior, fake reviews or astroturfing;
• generate content for political campaigning or lobbying that is deceptive about its origin;
• build systems that deceive people into believing they are interacting with a human when they are not, without clear disclosure.

1.4 High-risk and regulated activities

Do not use Output as the sole basis for decisions in the following areas without qualified human review and compliance with applicable law:

• medical diagnosis or treatment, or the operation of medical devices;
• legal, financial, insurance or tax advice to third parties;
• credit, employment, housing, education or benefits eligibility decisions;
• law enforcement, surveillance, biometric identification or predictive policing;
• operation of critical infrastructure, weapons, aviation, or other safety-critical systems.

1.5 Abuse the Service

• circumvent, disable or interfere with usage limits, credit accounting, rate limits, safety filters or security features;
• share accounts or API keys between unrelated parties, or resell access without a written reseller agreement;
• use the Service to build, train or improve a competing foundation model, or systematically extract Output for that purpose;
• reverse engineer, decompile or attempt to discover the source code, model weights or prompts of the Service, except where permitted by law;
• probe or test the vulnerability of the Service without authorization (see our responsible disclosure program on the Security page);
• send excessive requests that degrade the Service for others;
• upload content containing sensitive personal data (health, biometric, financial account, government ID numbers, data about children) except under a signed agreement that covers such data (available on Business and Scale plans).

1.6 Agent-specific rules

Agent mode can take actions in your workspace and sandbox. You may not direct the agent to:

• access systems, networks or data you are not authorized to access;
• exfiltrate secrets, credentials or personal data;
• perform destructive actions on production systems without explicit human confirmation;
• operate autonomously in a way that violates any part of this AUP.

You are responsible for supervising agent runs and for any outcome of applying agent output.

2. Content standards for applications built on the API

If you offer an application built on the Zyverna API to end users, you must:

• publish your own terms of service and privacy policy;
• implement appropriate content moderation and abuse reporting;
• not represent Output as human-authored where that would be misleading;
• comply with age restrictions and consumer protection law in your market;
• promptly respond to our requests regarding abuse originating from your application.

3. Enforcement

We may investigate suspected violations and take any action we consider appropriate, including issuing warnings, removing content, throttling, suspending or terminating accounts, withholding refunds to the extent permitted by law, and reporting unlawful activity to authorities. Where practicable we will notify you and give you an opportunity to respond, unless doing so would compromise an investigation or safety.

Automated systems may flag content for review. We do not use Customer Content to train models, including for safety classifiers.

4. Reporting abuse

To report a violation of this AUP, email [email protected] with a description and any supporting evidence. To report a security vulnerability, email [email protected].

5. Changes

We may update this AUP as the Service and legal landscape evolve. Material changes will be announced with at least 30 days' notice where practicable.