Terms of Service · RootInference

1. Acceptance of terms

This section contains placeholder content prepared for legal review. It describes the intended scope and structure of the final document and must be reviewed and approved by qualified counsel before publication.

2. The service

RootInference operates an inference gateway: a unified, OpenAI-compatible API that routes customer requests to third-party model providers (including, among others, OpenAI, Anthropic, Google, Meta and Mistral) and returns their responses. RootInference does not itself train or host foundation models; model outputs are generated by the selected third-party provider. The service additionally provides request routing, automatic failover between providers, usage analytics and credit-based billing. Model availability, behaviour and pricing may change as third-party providers change their offerings. This section is placeholder content prepared for legal review and must be approved by qualified counsel before publication.

3. Accounts and organisations

This section contains placeholder content prepared for legal review. It describes the intended scope and structure of the final document and must be reviewed and approved by qualified counsel before publication.

4. Credits, subscriptions and payment

Usage of the service is metered in credits, where 1 credit corresponds to approximately $0.01 of inference. Credits purchased as one-time packs remain valid for 12 months from the date of purchase. Credits included with a subscription roll over and remain valid for 12 months from the date they are granted, subject to an active subscription at the time of grant. Credits are consumed on a first-expiring, first-consumed (FIFO by expiry) basis. Credits are redeemable against usage of any model available through the platform, including models from OpenAI, Anthropic, Google, Meta and other providers. Except where required by law or set out in the Refund & Credit Policy, credits are non-transferable and not redeemable for cash. This section is placeholder content prepared for legal review and must be approved by qualified counsel before publication.

5. Customer data and intellectual property

Customers retain all rights in prompts submitted to the service and, to the extent permitted by the relevant model provider's terms, in model outputs returned to them. RootInference does not use customer prompts or outputs to train models, and passes data to model providers solely to fulfil the customer's requests. This section is placeholder content prepared for legal review and must be approved by qualified counsel before publication.

6. Acceptable use

This section contains placeholder content prepared for legal review. It describes the intended scope and structure of the final document and must be reviewed and approved by qualified counsel before publication.

7. Warranties and disclaimers

This section contains placeholder content prepared for legal review. It describes the intended scope and structure of the final document and must be reviewed and approved by qualified counsel before publication.

8. Limitation of liability

This section contains placeholder content prepared for legal review. It describes the intended scope and structure of the final document and must be reviewed and approved by qualified counsel before publication.

9. Termination

This section contains placeholder content prepared for legal review. It describes the intended scope and structure of the final document and must be reviewed and approved by qualified counsel before publication.

10. Governing law

This section contains placeholder content prepared for legal review. It describes the intended scope and structure of the final document and must be reviewed and approved by qualified counsel before publication.