Dieses Dokument wird auf Englisch bereitgestellt. Sollte eine Übersetzung bereitgestellt werden, ist die englische Fassung maßgeblich.

Acceptable Use Policy

Effective date: 6 October 2026

This Acceptable Use Policy applies to the Zelvimo AI API Service operated by Zelvimo OÜ, an Estonian private limited company with registry code 17582420 and registered office at Harju maakond, Tallinn, Haabersti linnaosa, Paldiski mnt 199a-4, 13517, Estonia ("we", "us" or "our"). It forms part of our Terms of Service and applies to all users, including use through products built on the Service.

1. Applicable provider and model rules

Requests are processed by the inference provider operating the model you select. The current inference provider is DeepInfra. You must comply with its applicable Terms of Service, as well as the licence, usage restrictions and policies that govern your selected model and are identified in the Service or model documentation.

A model developer and an inference provider may be different companies. A policy applies only where it actually governs the selected model or its delivery through the Service. Where an applicable provider or model rule is stricter than this Policy, that rule applies to requests using the relevant model, subject to mandatory law.

Provider rules do not displace our obligations under Estonian or EU law or your statutory consumer rights.

2. Prohibited uses

You must not use the Service to create, request or distribute content, or carry out activities, that fall within the following categories.

2.1 Harm to people

  • Sexualising or exploiting children, including creating or distributing child sexual abuse material. We report such material where required by law.

  • Promoting, inciting or facilitating violence, terrorism or violent extremism.

  • Threatening, harassing, bullying, stalking or intimidating anyone.

  • Encouraging self-harm or suicide, or providing instructions intended to facilitate it.

  • Helping develop or deploy biological, chemical, nuclear or radiological weapons, explosives or other weapons capable of mass harm.

  • Facilitating human trafficking or exploitation.

2.2 Illegal activities and violations of rights

  • Activities illegal where they occur or where the content is used.

  • Infringing intellectual property, privacy, confidentiality or other rights.

  • Collecting, processing or disclosing personal data without a lawful basis, including profiling without legal justification or processing special-category or criminal-offence data without the safeguards required by law.

  • Facilitating fraud, scams, phishing, money laundering or other financial crime.

  • Breaching sanctions or export-control measures applicable in Estonia, including applicable EU restrictive measures, or using the Service from a restricted location or for a prohibited person as described in the Terms.

2.3 Deception and manipulation

  • Fraudulently impersonating a person or organisation, or creating deceptive deepfakes of real people without their consent.

  • Running disinformation or influence campaigns through fake reviews, accounts or coordinated inauthentic behaviour.

  • Deploying an AI system to interact with people without the disclosure required by law, including the applicable transparency duties in Article 50 of Regulation (EU) 2024/1689 (the EU AI Act).

  • Presenting AI-generated content as human-created where this misleadingly causes harm, or failing to disclose synthetic or manipulated content where the law requires it.

  • Using manipulative or exploitative techniques likely to cause legally relevant harm, or carrying out any prohibited AI practice under Article 5 of the EU AI Act, such as prohibited social scoring or exploitation of vulnerabilities.

2.4 Decisions requiring safeguards

  • Making automated decisions that legally or significantly affect people, including decisions concerning credit, employment, housing, insurance, education, essential services or legal status, without the required lawful basis, safeguards and human oversight.

  • Deploying a high-risk AI system without satisfying the applicable legal requirements for your role and intended use. Access to an API does not establish that a particular application complies with those requirements.

  • Giving medical, legal or financial advice to others without qualified professional review where such review is required by law or the applicable provider policy.

2.5 Attacks and abuse of the Service

  • Creating or distributing malware, ransomware or tools intended for unauthorised access, or attacking systems or networks.

  • Testing the security of the Service without our written permission. Report suspected vulnerabilities to [email protected].

  • Overloading the Service or bypassing rate limits, spending limits, billing, content filters or security and safety measures.

  • Attempting to bypass model safeguards to obtain content prohibited by this Policy.

  • Sharing API keys publicly, selling or renting accounts or keys, or creating multiple accounts to obtain welcome credit.

  • Sending spam or bulk unsolicited messages.

  • Using outputs to develop competing models where the terms actually governing the selected model prohibit that use.

3. Responsibilities when building products

If other people use a product you build using Zelvimo, you must:

  • ensure its use complies with this Policy and impose corresponding rules on your users;

  • provide an abuse-reporting route and act on substantiated reports;

  • provide legally required privacy information and AI disclosures;

  • establish a lawful basis for personal data submitted through the Service, minimise the data and obtain required permissions;

  • put an Article 28 GDPR data processing agreement and any necessary transfer safeguards in place before submitting personal data for processing on your behalf; and

  • meet the AI Act and other legal obligations applicable to your role and the product's actual use.

You may build applications that use the Service within the Terms. You may not resell the Service itself or provide raw, standalone gateway access unless we have agreed otherwise in writing.

4. How we address breaches

We do not routinely read request content or pre-screen every request. We may act after a report, an abnormal pattern in request metadata, or a notification from a provider, payment service provider or competent authority. Any access to personal data for an investigation must have a lawful basis and be limited to what is necessary.

Depending on the evidence, urgency and seriousness of the issue, we may:

  • warn you and require corrective action;

  • block particular requests, models or API keys;

  • suspend or close an account in accordance with section 11 of the Terms; or

  • make a report to a competent authority where required or lawfully justified.

Measures must be proportionate. Where lawful and practicable, we will explain the reason and give an opportunity to remedy the issue. An urgent security or legal restriction may be applied immediately. A suspension or closure does not automatically forfeit unused paid balance; the Terms and Refund Policy govern any refund or lawful withholding.

You may challenge a decision by emailing [email protected] with your account email, the decision and relevant supporting information. We review challenges according to urgency and available evidence. Consumer complaints are answered within the period in section 18 of the Terms. We may withhold details where disclosure would unlawfully reveal personal data, compromise security or an investigation, or breach a legal duty. Your statutory rights and access to competent authorities or courts remain unaffected.

5. Reporting abuse

Email [email protected] to report abuse or illegal content. Identify the relevant request, account or location where possible, explain the issue and include supporting information you are entitled to share. Do not send live API keys, passwords or unnecessary sensitive personal data.

6. Changes and applicable law

Changes to this Policy follow the notice and amendment rules in section 17 of the Terms. This Policy is governed by the laws of Estonia and applicable EU law. Mandatory consumer and data protection rights remain unaffected.

7. Contact

Zelvimo OÜ
Operator of Zelvimo
Registry code: 17582420
Registered office: Harju maakond, Tallinn, Haabersti linnaosa, Paldiski mnt 199a-4, 13517, Estonia
Abuse reports and enquiries: [email protected]