Acceptable Use Policy

1. Purpose and Scope

This Acceptable Use Policy ("AUP") forms part of the Terms of Service and any agreement that references it, including any licence agreement for self-hosted software. It applies to customers, their Users and anyone accessing the Services. Customers are responsible for ensuring that their Users and the agents they configure comply with it.

2. Prohibited Uses: Illegal or Harmful Activity

You may not use the Services to:

• violate any applicable law or regulation, or infringe or misappropriate anyone's intellectual property, privacy or other rights;
• process personal data without a lawful basis, required notices or consents, or in breach of data protection law;
• engage in fraud, phishing, deceptive practices, impersonation or the creation of fake reviews or identities;
• harass, threaten, defame, stalk or discriminate against any person, or generate content that sexualizes minors;
• develop or support weapons capable of mass casualties, including chemical, biological, radiological or nuclear weapons;
• facilitate terrorism, violent extremism, human trafficking or any other serious harm;
• violate export control or sanctions laws, or provide access to embargoed countries or sanctioned persons.

3. Prohibited Uses: Security and Platform Integrity

You may not:

• access, or attempt to access, systems, accounts or data you are not authorized to use, including through connectors;
• probe, scan or test the vulnerability of the Services without our written permission (please report vulnerabilities to admin@vouchstone.ai);
• create, distribute or deploy malware, ransomware or any code designed to disrupt, damage or gain unauthorized access;
• circumvent guardrails, autonomy levels, human-approval gates, access scopes, rate limits, token budgets or audit logging;
• tamper with, delete or falsify entries in the signed audit ledger;
• overload the Services, conduct denial-of-service attacks or interfere with other customers;
• share credentials or seats, or resell access except as permitted in an Order Form;
• use the Services or Output to develop a competing product or to extract model weights, prompts or training data.

4. Responsible Agent Deployment

Because Vouchstone agents can take actions in connected systems, you must:

• configure agents with the minimum permissions and data scopes they need;
• keep appropriate human oversight and approval for actions that could significantly affect individuals, finances, safety or legal rights;
• not deploy agents that make solely automated decisions producing legal or similarly significant effects on individuals (such as employment, credit, insurance, housing, education or healthcare decisions) unless you have implemented the human review, notices and safeguards required by law;
• not use the Services for social scoring, manipulative techniques exploiting vulnerabilities, untargeted scraping of facial images, or emotion recognition in workplaces or schools;
• disclose to people that they are interacting with an AI system where the law requires it;
• monitor agent behaviour and promptly stop any agent acting outside its intended purpose.

5. Regulated and Sensitive Data

You may not submit protected health information, payment card data, government identification numbers, biometric data or other special categories of data unless your Order Form permits it and all required agreements (such as a Business Associate Agreement) and safeguards are in place. Self-hosted and air-gapped deployments are recommended for the most sensitive workloads.

6. Third-Party Models and Systems

When the Services route requests to third-party AI models or connected systems, you must also comply with those providers' usage policies, including the usage policies of Anthropic, OpenAI and any other model provider used for your workloads.

7. Enforcement

We may investigate suspected violations and may remove content, disable agents or connectors, suspend or terminate accounts, and cooperate with law enforcement. Where practicable, we will notify you and give you an opportunity to fix the issue, but we may act immediately where there is a risk of harm, a security threat or a legal requirement.

8. Reporting

Report violations or abuse to legal@vouchstone.ai. Report security vulnerabilities to admin@vouchstone.ai with the subject line "Security." We welcome good-faith security research that follows responsible disclosure and does not access or alter other customers' data.

Vouchstone LLC, 2000 Walnut Ave, Fremont, California 94538, United States