Terms of Service
Effective 2026-08-27 · v1.0
1. Service
Inference Yield provides an OpenAI-compatible API for hosted open-weight language models. The service is provided on an as-available basis during our pilot phase; current availability is published on the status page.
2. Acceptable use
You may not use the service to violate applicable law, to attempt to extract other customers' data, to probe or disrupt the infrastructure, or to generate content that violates the upstream model's license terms (Apache 2.0 for Qwen3.8 weights). We may rate-limit or suspend keys engaged in abuse.
3. Billing
Usage is billed per token at the prices published on the home page and in /v1/models, with cached input tokens billed at the cached rate. Metering records (metadata only — see Privacy) are the billing source of truth.
4. Data
Prompt and completion content is not retained. See the Privacy Policy, which is part of these terms.
5. Warranty & liability
The service is provided "as is" without warranty of any kind. To the maximum extent permitted by law, our aggregate liability is limited to the fees you paid in the month the claim arose. Model outputs are generated by open-weight models and are your responsibility to review before use.
6. Changes
We may update these terms with notice on this page. Continued use after the effective date constitutes acceptance.