ModelStation

ModelStation Acceptable Use Policy

Operator: ELVARON LIMITED, Registration No. 79402144, 14/F, China Building, 29 Queen's Road Central, Central, Hong Kong. Contact for reports of violations: support@modelstation.org

Version: 2.2 Effective date: from the date of publication on the website.

1. Purpose and scope

1.1. This Policy protects users, AI Providers, third parties, and the stability of the Service. It applies to the website, chat, agents, memory, file attachments, automation, administrative interfaces, and programmatic access, if provided.

1.2. This Policy forms part of the Terms of Use. A breach of it is a breach of contract.

1.3. This Policy applies in addition to the Model Provider Terms. If a provider’s rules are stricter, the stricter rules apply. You must know and comply with the rules of the provider of the Model you use.

1.4. The requirements apply to you, your employees, contractors, invited users of a corporate account, and your own customers if you provide them with access to the results of the Service. You are responsible for their actions as for your own.

2. Lawfulness and third-party rights

It is prohibited to use the Service to:

  1. violate the applicable law of any affected jurisdiction;
  2. violate sanctions, export, and import restrictions, as well as AI Provider restrictions by territory and category of persons;
  3. infringe intellectual-property rights, including reproduction, adaptation, and circumvention of technical protection measures for works;
  4. violate the right to privacy, secrecy of correspondence, trade secrets, medical, banking, and attorney-client privilege;
  5. breach contractual obligations to third parties, including confidentiality obligations;
  6. circumvent blocks imposed by the Company, a provider, or a competent authority.

3. Prohibited content and purposes

3.1. Absolutely prohibited (a violation results in immediate blocking without warning and, where there is a legal duty, a report to competent authorities):

  1. harmful content generation in any form: text, images, audio, video, code, and any other materials;
  2. sexual materials involving minors, their description, stylization, or generation in any form;
  3. NSFW, explicit, and sexually suggestive generation: pornography, erotica, sexually explicit or sexually suggestive content in any form, including text, images, audio, and video, including content involving adults;
  4. materials directed at sexual exploitation, human trafficking, forced labor;
  5. preparation of or assistance to terrorism, violent extremism, mass violence;
  6. instructions for creating weapons, explosives, chemical, biological, radiological, or nuclear agents;
  7. development of malicious software, exploits, ransomware, tools for circumventing protection systems;
  8. sexualized content involving real people without their consent, including content created on the basis of their images;
  9. content depicting sexual violence;
  10. extreme violence, gore, dismemberment;
  11. zoophilia;
  12. deepfakes without the consent of the people depicted;
  13. hate speech.

3.2. The Service is not intended for harmful content generation or NSFW generation. Requesting, submitting, or generating harmful content, or explicit or sexually suggestive materials, in any form is prohibited.

3.3. Prohibited without a lawful basis and appropriate control measures:

  1. creation of misleading synthetic content (deepfakes) with the image, voice, or signature of a real person;
  2. manufacture or imitation of documents, identity documents, seals, signatures, payment documents;
  3. impersonation of another person, organization, government authority, or of the operator of the Service itself;
  4. defamation, dissemination of knowingly false information, bullying, harassment, threats, incitement of hatred, and discrimination on protected characteristics;
  5. fraud, phishing, social engineering, get-rich-quick schemes, market manipulation, circumvention of KYC and anti-money-laundering procedures;
  6. mass unsolicited communications, spam, inflation of reviews, ratings, and engagement, automated creation of fictitious identities;
  7. unlawful surveillance, stalking, de-anonymization, facial recognition, and profiling of people without a lawful basis;
  8. activity requiring a license, in the absence of such a license on your part (financial, medical, legal services, and the like).

4. Security and integrity of the Service

It is prohibited to:

  1. conduct unauthorized scanning, penetration testing, load testing, and vulnerability searching without written approval from the Company;
  2. exploit vulnerabilities, bypass authentication, escalate privileges, access others’ accounts, data, and administrative interfaces;
  3. circumvent limits, quotas, filters, security checks, billing restrictions, and attempts to obtain the service without a Balance debit;
  4. extract system instructions and service parameters of the Service, reverse-engineer, decompile, and create derivative solutions on their basis;
  5. scrape, automatically collect data, copy the model catalog, prices, and interfaces to create a competing product;
  6. distribute malicious code, interfere with the operation of networks and servers, conduct availability attacks, create excessive load;
  7. use the Service as an open proxy or relay for third-party access.

5. Adversarial testing of models (Red Teaming)

5.1. Red Teaming — prompt injection, jailbreak, attempts to bypass a model’s protective mechanisms, extract training data or system instructions, and other adversarial impacts.

5.2. Such actions violate the rules of most AI Providers and are prohibited without the Company’s prior written permission.

5.3. A request for research testing is sent to support@modelstation.org with a description of the purpose, methodology, scope, protective measures, and the procedure for disclosing results. Permission is granted at the Company’s discretion, in individual cases and for a limited period.

5.4. Uncoordinated attempts are recorded by the monitoring systems of the Company and providers and generally result in termination of access, including on the provider’s side.

6. Automation, limits, and fair use

6.1. Automated access is permitted only through the provided interfaces and within reasonable limits.

6.2. It is prohibited: to distribute traffic across accounts, keys, IP addresses, and proxies for the purpose of circumventing limits; to create multiple accounts by one person to obtain bonuses, circumvent blocks or limits; to launch uncontrolled loops and retries without exponential backoff; to resell access and provide access to third parties without the Company’s written consent.

6.3. The Company applies technical limits on the frequency and parallelism of requests and may change them without prior notice to protect the Service.

6.4. The User must control the User’s own consumption: expenses arising from the User’s automation, integration errors, or suboptimal requests are borne by the User.

7. Sensitive data and use-case restrictions

7.1. The User must have a lawful basis for transmitting to the Service any personal data and must provide the necessary notices and consents.

7.2. The Service is not intended for use as a specialized system for processing or storing: health data and medical records; payment details and data subject to PCI DSS; biometric data; government identifiers; children’s data; information classified as a state secret or subject to a special protection regime.

7.3. It is prohibited to use Outputs as the sole or determining basis for decisions concerning: medical care and treatment; lending, insurance, and financial access; hiring, dismissal, evaluation, and monitoring of employees; admission to education and examination assessment; the provision of public and social services; legal status, migration, deportation; the application of sanctions and coercive measures; life safety and the management of critical infrastructure.

7.4. If your use case is classified by applicable artificial-intelligence law as high-risk or subject to special regulation, you act as an independent responsible person and must ensure risk assessment, human involvement, documentation, information of affected persons, and labeling of synthetic content.

7.5. Practices expressly prohibited by applicable artificial-intelligence legislation are prohibited, including manipulative techniques exploiting people’s vulnerabilities, social scoring, and prohibited biometric categorization.

8. Requirements for corporate accounts

8.1. The administrator of a corporate account is responsible for: the composition of invited users; settings for history retention, memory, and logging; compliance of use with the organization’s internal policies and applicable law.

8.2. The organization must communicate the requirements of this Policy to its users and ensure their compliance.

9. Control and response measures

9.1. The Company is not obliged to conduct prior review of the contents of requests, but may apply automatic and manual checks where there are grounds to suspect a violation.

9.2. Depending on the nature and seriousness of the violation, the following are applied:

  1. a warning and a request for explanations;
  2. restriction of request frequency, quotas, or individual Models;
  3. rejection of specific requests;
  4. disabling of a feature, suspension of a key or account;
  5. deletion or blocking of content;
  6. termination of the contract and deletion of the account;
  7. notification of the AI Provider, payment partner;
  8. notification of competent authorities, if required or permitted by law;
  9. assertion of claims for damages.

9.3. In the event of an immediate threat to security, third-party rights, or the stability of the Service, measures are applied immediately and without prior notice.

9.4. Upon termination of access due to a violation, unused Balance is not refunded to the extent permitted by law. The Company is not obliged to retain or transmit deleted content.

9.5. Decisions on the application of measures may be appealed by contacting support@modelstation.org. We review the communication and, where necessary, provide human review of the decision. The Company retains the right of final decision to the extent established by law.

10. Reporting a violation

10.1. Reports of abuse should be sent to support@modelstation.org, specifying: date and time; a description of the violation; links, request identifiers, or other available indicators; your contact details; and, for a rights-infringement report — confirmation of rights and a statement of good faith.

10.2. Knowingly false reports directed at restricting another user’s access constitute a violation of this Policy.

11. Changes

The Policy may be updated in connection with changes in legislation, AI Provider requirements, and the emergence of new risks. Material changes are introduced in accordance with §21 of the Terms of Use; changes directed at eliminating an immediate threat apply from the moment of publication.