1. Scope and Services
1.1 These Special Terms and Conditions supplement the General Terms and Conditions (GTC) of Hetzner Online GmbH (hereinafter “we” or “us”). They apply to all contracts for AI services with business customers (§14 of the German Civil Code (BGB)) and consumers (§13 of the German Civil Code (BGB)) (hereinafter “Customer”).
1.2 AI services consist of access to AI models, application programming interfaces (“APIs”), and other AI-based functions. The Customer uses these services to process inputs (“Input”) and receive automatically generated results (“Output”). The nature, scope, performance characteristics, availability, and usage limits are set forth in the respective service description.
1.3 The AI Services may be based on models and services provided by third parties. To the extent that their terms of use apply, they will be made available to the Customer prior to or at the time of use. The Customer must then comply with them. In the event of a material breach, we may restrict the affected service in accordance with Section 3.6.
2. Characteristics of the Output, Testing Requirements
2.1 The output is based on statistical methods. It may be incorrect, incomplete, outdated, or unsuitable, and may cite non-existent sources. We do not guarantee that the output is accurate, complete, up-to-date, original, protectable, free from third-party rights, or suitable for a specific purpose, unless this has been expressly agreed upon.
2.2 The customer is solely responsible for reviewing the output before using it, publishing it, or basing decisions on it. For high-risk applications (particularly those involving medical, legal, tax, financial, or security-critical matters), an expert review is required. The AI services do not replace qualified professional advice.
3. Customer Obligations and Permitted Use
3.1 The customer shall ensure that it is legally permitted to process and transmit the input to us and that this does not infringe any third-party rights (in particular, copyrights, trademark rights, personality rights, and data protection rights). This also applies to any use of the output.
3.2 The AI services may only be used within the scope of the contract and in accordance with applicable law. In particular, but not limited to, the following uses are prohibited: the use of the services to create illegal, pornographic, or obscene materials; extremist content or content that violates public decency; material that is likely to seriously endanger the moral well-being of children or adolescents; or material that infringes the rights of third parties (in particular copyrights, rights to a name, trademark rights, and data protection rights). This includes the creation of defamatory content, insults, or the denigration of individuals or groups of people.
3.3 The following, in particular, constitute misuse:
- significantly exceeding the usage limits set forth in Section 3.3, for example through the use of scripts, macros, RPA tools, or other automation; the mass insertion of pre-generated content; running parallel sessions to circumvent limits;
- the transfer of the AI Services to third parties;
- the circumvention of technical usage limits;
- the use of multiple accounts to circumvent limits or protective measures;
- the impairment of the security, integrity, or availability of the AI Services or our infrastructure, such as through overloading, unauthorized security tests, or attacks.
3.4 Use that is expressly provided for in the Service Description or has been approved by us is not considered abusive; this includes, in particular, automated use via APIs, MCP servers, or the integration of our own or third-party AI models. If a connected model acts on behalf of the customer using the customer’s authorizations, this does not constitute a transfer to third parties. Sections 3.4(c) through (e) remain unaffected.
3.5 We may impose technically reasonable limits on the number, frequency, and concurrency of requests to ensure stability, security, and fair resource allocation (rate limiting).
3.6 In the event of concrete, objectively verifiable indications of misuse or use in violation of the contract, we may take proportionate temporary measures. These include throttling requests, restricting or blocking individual functions, models, or access, as well as switching to another available model. We will inform the customer of the measure and the reasons for it and give the customer the opportunity to comment and remedy the situation. This does not apply to the extent that doing so would jeopardize threat prevention, investigations, or legitimate security interests. We will lift the measures as soon as the reason for them no longer applies.
4. Rights and Use of Data
4.1 We use input and output exclusively
- to provide the AI services,
- to ensure functionality, stability, and security, as well as for troubleshooting,
- for SaaS services, to improve technical performance and quality,
- for SaaS services, to prevent misuse, provided there is an objective occasion or concrete suspicion in individual cases.
4.2 The customer is solely responsible for data that it transmits directly to third-party providers via a designated connection, to the extent that such processing occurs outside our control. Responsibilities under data protection law, any data processing agreements, and statutory data protection rights remain unaffected.
5. Connected AI Models and Automated Actions
5.1 To the extent that the Customer integrates its own or third-party AI models, agents, or automated systems, these may perform actions—including erroneous or unintended ones—without individual authorization.
5.2 The Customer is responsible for actions based on its inputs, instructions, configurations, or authorizations, particularly if they are performed using its login credentials or API keys. The customer is solely liable to us for all costs incurred as a result of these actions. This applies regardless of whether the customer performed the actions themselves or whether third parties using their services did so. The customer may not hold us liable for these third parties. Any claims the customer may have against these third parties remain unaffected.
The Customer is solely responsible for limiting permissions to what is necessary and for implementing appropriate safeguards such as approval processes, logging, and backup copies.
5.3 Our liability for our own culpable breaches of duty continues to be governed by Section 9 of our General Terms and Conditions.
6. Modifications to Models and Components
6.1 We may modify, replace, or discontinue AI models, components, and interfaces for objective reasons, in particular due to technical advancements, security requirements, legal requirements, or changes in the availability or terms of third-party providers. §327r of the German Civil Code (BGB) applies to consumers. We will provide timely notice. In the event of more than a minor disruption, statutory rights apply, in particular the right to terminate the contract without penalty.
6.2 If compelling technical, security-related, or legal reasons require immediate action, we may take temporary measures without prior notice and will inform the customer immediately thereafter.
7. Availability
The performance of the AI services may depend on third-party providers. Fluctuating response times and temporary restrictions may occur. In the event of a failure of a third-party service, we may implement appropriate alternative solutions.
8. AI Regulation (EU) 2024/1689
8.1 Each party shall fulfill the obligations under the AI Regulation that apply to it based on its actual role. For example, if the customer is the operator of an AI system, it is responsible for the corresponding obligations.
8.2 The Customer may be considered a provider, in particular if the Customer places an AI system on the market under its own name or brand, substantially modifies a high-risk AI system, or changes its intended purpose in such a way that a high-risk AI system is created.
8.3 The use of the AI Services for high-risk AI systems is permitted only with our prior written consent. There is no entitlement to such consent. The Customer shall determine the classification of its use on its own. Our consent does not release the Customer from its legal obligations.
9. Indemnification
9.1 The Customer shall indemnify us against any third-party claims arising from a breach of statutory or contractual obligations in connection with the AI Services for which the Customer is responsible. This includes, in particular, unlawful input, the unlawful use of output, and actions taken by connected systems as described in Section 5. The indemnification shall cover the reasonable costs of legal defense. Any contributory negligence on our part shall reduce the indemnification accordingly.
9.2 We shall immediately inform the Customer of any claims asserted and, to the extent possible, allow the Customer to participate in the defense. The Customer may not acknowledge any claims or enter into any settlements at our expense without our consent.
10. Final Provisions
10.1 Our General Terms and Conditions apply in addition to these provisions. In the event of any conflict, these Special Terms and Conditions for AI Services shall take precedence.
10.2 If any provision is invalid, the remaining provisions shall remain valid. The invalid provision shall be replaced by the applicable statutory provision.
Effective as of: November 17, 2026