Acceptable Use Policy

Effective date: February 27, 2026
Last updated: August 23, 2026

This Acceptable Use Policy (“Policy”) explains prohibited and restricted conduct when using Magister AI’s Service. It is designed to protect people, customer data, third-party rights, and the security and reliability of the Service without unnecessarily restricting legitimate research, analysis, creativity, or discussion.

This Policy forms part of the agreement that governs your use of the Service. Capitalized terms not defined here have the meanings given in the applicable Consumer Terms, business agreement, or other agreement between you and Magister AI LLC (“Magister AI,” “we,” “us,” or “our”).

1. Scope and responsibility

This Policy applies to everyone who accesses or uses the Service, including through an account provided by an organization. Magister AI calls the AI assistants and tool-using features available through the Service “Companions.” Companions may generate Output and, where enabled, take Actions such as browsing, executing code, changing files, or interacting with connected systems.

This Policy applies to Inputs, Outputs, Companions, Actions, integrations, automation, shared Content, and use of third-party models or services through Magister AI.

You must not use the Service, or help another person use the Service, to engage in the conduct prohibited below. You also must not deliberately structure a request, workflow, account, or integration to evade this Policy.

If you allow another person to use the Service through your account or organization, you are responsible for taking reasonable steps to ensure that they know and follow this Policy.

1.1 Context matters

We assess use based on relevant context, including apparent intent, authorization, capabilities requested, pattern of activity, and reasonably foreseeable effects. A topic is not prohibited merely because it is sensitive, controversial, fictional, or capable of misuse.

Good-faith research, education, journalism, commentary, artistic or fictional work, safety analysis, and authorized security testing are not prohibited solely because they discuss or examine harmful conduct. Those uses must still comply with this Policy and applicable law.

1.2 Third-party models

The Service may provide access to models supplied by third parties. A model provider may impose additional restrictions on use of its model. Where provider-specific restrictions are identified in the Service or applicable documentation, they apply to your use of that model. A provider may refuse or limit a request under its own rules.

Provider-specific restrictions do not become general restrictions on your use of other parts of the Service unless we expressly state otherwise.

2. Unlawful conduct and violations of rights

You must not use the Service to:

  1. 1. engage in, facilitate, or materially assist conduct that violates applicable law, a court order, sanctions, or export-control restrictions;
  2. 2. offer, obtain, or distribute illegal goods or services;
  3. 3. infringe, misappropriate, or violate another person’s intellectual-property, privacy, publicity, confidentiality, contractual, or other legal rights;
  4. 4. upload, disclose, or use Content that you do not have the right or authority to use;
  5. 5. disclose trade secrets, confidential information, or legally protected records without authorization; or
  6. 6. conceal or launder the proceeds of unlawful activity.

A good-faith request to understand, detect, prevent, document, or report unlawful conduct is not prohibited merely because it discusses that conduct.

3. Abuse, exploitation, and serious harm

You must not use the Service to:

  1. 1. make credible threats, incite violence, coordinate targeted harassment, stalk, dox, extort, coerce, or intimidate another person;
  2. 2. facilitate human trafficking, sexual exploitation, kidnapping, or other violent abuse;
  3. 3. create, obtain, sexualize, distribute, or facilitate child sexual abuse material or the grooming or sexual exploitation of a minor, including where the material is wholly or partly generated;
  4. 4. create or distribute non-consensual intimate imagery or deceptively place a real person’s likeness in a sexual or intimate context without authorization;
  5. 5. encourage, instruct, or materially facilitate another person’s suicide, self-harm, or disordered eating;
  6. 6. materially facilitate a terrorist act, violent criminal act, or the deliberate infliction of serious physical harm;
  7. 7. design, construct, acquire, or deploy biological, chemical, radiological, or nuclear weapons or other weapons of mass destruction; or
  8. 8. deliberately damage or destroy property without authorization.

This section does not prohibit consensual, lawful fiction or art; historical, academic, medical, or journalistic discussion; harm-prevention work; or emergency and support resources, provided the use does not facilitate prohibited harm.

4. Fraud, deception, impersonation, and manipulation

You must not use the Service to:

  1. 1. commit or facilitate fraud, scams, phishing, identity theft, credential theft, payment abuse, money laundering, or deceptive financial schemes;
  2. 2. forge or materially falsify identity documents, government records, financial records, signatures, credentials, evidence, or other documents for a deceptive or unlawful purpose;
  3. 3. impersonate a real person or organization, or use a person’s likeness or voice, in a manner reasonably likely to deceive others about identity, authority, endorsement, or authenticity;
  4. 4. distribute spam, fake reviews, fabricated endorsements, or coordinated inauthentic activity intended to mislead people about its source, scale, or support;
  5. 5. conceal the artificial origin of Content where disclosure is required by law or reasonably necessary to avoid material deception;
  6. 6. misrepresent your relationship with Magister AI or make a service appear to be operated, sponsored, or endorsed by us when it is not; or
  7. 7. deceive people about voting eligibility, voting procedures, election dates or locations, or otherwise interfere unlawfully with voting or election systems.

Parody, satire, fiction, pseudonymous expression, simulation, and authorized testing are not prohibited merely because they involve an assumed identity, provided the context does not create material deception, rights violations, or another prohibited harm.

5. Privacy, personal data, and surveillance

You must not use the Service to:

  1. 1. obtain, aggregate, infer, expose, sell, or distribute personal data without authorization or another lawful basis where one is required;
  2. 2. dox a person or expose private contact details, credentials, health information, financial information, precise location, communications, or other sensitive information in order to harm, exploit, intimidate, or unlawfully monitor them;
  3. 3. track, surveil, or intercept a person’s location, device, communications, or activity without lawful authority;
  4. 4. create or use facial-recognition, biometric-identification, emotion-inference, or sensitive-trait-inference systems in violation of applicable law or without required notice, consent, and safeguards; or
  5. 5. obtain or use authentication secrets, session tokens, payment credentials, or other access credentials without authorization.

You remain responsible for providing required notices, obtaining required consents, and establishing a lawful basis for personal-data processing.

6. Cybersecurity and malicious code

You must not use the Service to:

  1. 1. gain or maintain unauthorized access to an account, device, application, network, data store, or other system;
  2. 2. discover, test, or exploit vulnerabilities without authorization from the owner of the relevant system;
  3. 3. create, deploy, distribute, or conceal malware, ransomware, destructive code, credential-stealing tools, botnets, or persistent unauthorized access tools;
  4. 4. conduct denial-of-service activity, destructive interference, unauthorized cryptomining, data exfiltration, or large-scale credential attacks;
  5. 5. evade endpoint protection, monitoring, authentication, or other security controls for an unauthorized or malicious purpose;
  6. 6. intercept communications or monitor devices without authorization; or
  7. 7. target the Service, its infrastructure, its providers, or other users for unsolicited vulnerability testing or exploitation.

Authorized defensive security work is permitted. This includes security administration, vulnerability remediation, malware analysis in a controlled environment, capture-the-flag exercises, education, and penetration testing performed on systems you own or are expressly authorized to test. You must remain within the scope of that authorization and use reasonable safeguards to avoid harm to other systems, data, or people.

7. Integrity and security of the Service

You must not:

  1. 1. bypass or interfere with authentication, safety controls, access restrictions, geographic restrictions, rate limits, usage limits, spending limits, or technical protections;
  2. 2. manipulate subscription, billing, metering, trial, referral, account-provisioning, or credit systems to avoid charges or obtain unauthorized access, refunds, or credits;
  3. 3. create or use accounts, devices, identities, or organizations to evade a restriction, suspension, usage limit, or enforcement action;
  4. 4. introduce malware, spam the Service, conduct denial-of-service activity, or otherwise degrade its operation or another user’s access;
  5. 5. probe, scan, test, or attempt to penetrate the Service without our prior written authorization;
  6. 6. reverse engineer, decompile, disassemble, extract, or attempt to discover non-public source code, model weights, algorithms, system prompts, or underlying components, except to the extent applicable law does not allow this restriction;
  7. 7. scrape, harvest, systematically extract, resell, or replicate Service data, Content, or model behaviour except through a feature we provide for that purpose or with our written authorization;
  8. 8. use the Service or systematically collected Output to build, train, fine-tune, evaluate, or improve a competing model or service without our written authorization;
  9. 9. remove or circumvent ownership, provenance, attribution, safety, watermark, or legal notices where present;
  10. 10. publish deceptive or materially misleading benchmarks, security findings, or test results, including by omitting material methodology; or
  11. 11. resell, sublicense, or make the Service available as a standalone substitute except under an agreement that permits it.

Good-faith evaluation and ordinary comparison of model quality are allowed. You must not evade controls, misrepresent methodology, expose protected information, or materially disrupt the Service while doing so.

8. Automated and bulk activity

Automated activity is not prohibited merely because it is automated. Automation initiated through supported Service features may be used if it complies with this Policy, remains within applicable limits, and does not degrade the Service or another user’s experience.

External automated access, including access through bots, scripts, or browser automation, is permitted only where we expressly provide or approve an interface for that use. You must not use automation to bypass intended controls, operate abusive account networks, conceal prohibited conduct, or consume disproportionate resources.

Bulk extraction, scraping, harvesting, mirroring, or replication of Service data or Output is prohibited unless we expressly authorize it in writing or provide a feature intended for that purpose.

9. Companions, Actions, and connected systems

You must not use a Companion or Action to:

  1. 1. access, alter, delete, transmit, publish, purchase, or communicate beyond the permissions you lawfully hold;
  2. 2. take an action on another person’s behalf without their authorization;
  3. 3. conceal material actions from a person whose account, property, data, or rights are affected;
  4. 4. deploy destructive or irreversible changes to a production system without appropriate review, authorization, backups, and recovery controls; or
  5. 5. evade a confirmation, permission, spending, or safety control.

You must scope permissions appropriately, use least-privilege access, review consequential Actions, and maintain reasonable safeguards and backups. You remain responsible for workflows and Actions you configure or authorize.

10. High-impact and safety-critical uses

You must not use a Companion, Output, or Action as the sole basis for a decision that has a legal or similarly significant effect on a person in areas such as credit, employment, education, housing, insurance, healthcare, legal services, or access to essential government or private services.

Where such use is lawful and otherwise permitted, you must provide qualified human review and any testing, documentation, notice, consent, explanation, security, and appeal process required by applicable law.

You must not use the Service as the sole control for a medical device, vehicle, weapon, critical-infrastructure system, emergency-response system, or other safety-critical system where failure could reasonably cause death, serious injury, or substantial property or environmental damage.

11. Sensitive and regulated data

Do not submit highly sensitive or specially regulated data unless the Service expressly states that it supports that data and all required agreements and safeguards are in place. Depending on context, this may include:

  • passwords, private keys, authentication secrets, or session tokens;
  • full payment-card or financial-account data;
  • protected health information;
  • government identification numbers;
  • biometric templates;
  • export-controlled or classified information; and
  • personal data about children.

This section does not prohibit using the Service to detect, remove, redact, secure, or respond to accidental exposure of such data, provided you are authorized to do so and use appropriate safeguards.

12. Reporting and cooperation

Please report suspected abuse, policy violations, compromised accounts, or security issues to [email protected]. Include enough information for us to understand and investigate the issue, but do not send passwords, private keys, full payment-card data, or unnecessary sensitive information.

If you discover a security vulnerability, please do not exploit it beyond what is reasonably necessary to confirm the issue, access another person’s data, degrade the Service, or publicly disclose details before we have had a reasonable opportunity to investigate and remediate it.

You agree to cooperate reasonably with an investigation into suspected misuse associated with your account, including by preserving relevant information and taking reasonable steps to secure the account. This does not require you to waive a legal right or privilege.

13. Enforcement

We may use automated signals and reviews conducted by artificial-intelligence systems and, where appropriate or required by law, human reviewers to detect and investigate suspected abuse, consistent with our Privacy Notice. A review may be performed by artificial intelligence, a human reviewer, or a combination of both. We do not undertake a general obligation to monitor all Content or activity.

Depending on the nature, severity, frequency, context, and urgency of a suspected violation, we may:

  • warn you or request corrective action;
  • block or remove particular Content or Actions;
  • limit a Companion, feature, model, integration, or workflow;
  • apply temporary rate limits or other protective measures;
  • suspend or terminate an account or organization;
  • preserve relevant information;
  • notify an organization administrator where appropriate; or
  • report conduct or information to authorities where required by law or reasonably necessary to address an imminent threat to rights or safety.

We may act without prior notice where reasonably necessary to stop active abuse, protect a person or system, preserve security, comply with law, or prevent evasion. In other cases, we will provide notice and an opportunity to correct the issue where appropriate.

If you believe we made an enforcement error, you may appeal by contacting [email protected]. Please identify the affected account, decision, and relevant context. Appeals may be reviewed by artificial-intelligence systems and, where appropriate or required by law, human reviewers. We will consider appeals reasonably and provide any explanation or remedy required by law.

14. Policy updates

We may update this Policy to reflect changes to the Service, law, security risks, abuse patterns, or our enforcement practices.

For a material change, we will provide reasonable advance notice through the Service, by email, or by another appropriate method, unless an urgent legal, safety, or security reason requires an earlier effective date. The date at the top identifies the latest revision.

If you do not agree to an updated Policy, you must stop using the Service. Continued use after the effective date of an update constitutes acceptance to the extent permitted by law.

15. Contact

Questions, reports, and appeals concerning this Policy may be sent to:

Magister AI LLC 2093 Philadelphia Pike, Suite #4099 Claymont, DE 19703 United States

Email: [email protected]