Annex B — AI Provider Annex (Provider Matrix; Pass-Through)

1. Integrated Providers (current adapters)

The Service may route requests to one or more AI Providers, which currently include: Anthropic, OpenAI, FireworksAI, Infercom, TogetherAI.

2. Pass-through terms and disclaimers

• Your use of a provider/model may be subject to provider terms, acceptable use policies, and data processing terms.
• Infere does not control, and is not responsible for, the operation, availability, legality, or compliance of any AI Provider. Each AI Provider acts as an independent third party. To the maximum extent permitted by law, Infere shall not be liable for any act or omission of an AI Provider, including any Output generated.
• Infere does not guarantee any provider's retention/training settings; those depend on provider terms and configuration.
• Providers may change APIs, prices, and features; Infere may update routing accordingly.
• Outputs may be inaccurate, incomplete, or infringe third-party rights. User is solely responsible for evaluating Outputs prior to reliance or use.

3. Output and IP

• Outputs are generated by providers; Infere provides no warranties regarding Outputs.
• Output ownership is allocated per Section 8 of the Terms, subject to Provider terms and third-party rights.

4. Provider-specific restrictions (placeholders)

In addition to these Terms, Users must comply with any mandatory provider-specific restrictions notified by Infere through the Service, documentation, or support channels, including geographic restrictions, restricted categories of content, prohibited use cases, training or retention constraints, and model-specific limitations.

5. Role clarification

The parties acknowledge that, unless otherwise expressly agreed, Infere acts as an intermediary service provider and not as an AI system provider within the meaning of Regulation (EU) 2024/1689 in respect of third-party models. The User remains responsible for the use of Outputs and for compliance with applicable AI governance obligations.