Subprocessors

Where a request is sent is decided by the inference endpoint that serves it, not by the developer of the model. Every model card and every response names the region and the sovereignty class of the endpoint that answered, and a routing policy that admits only EU classes will not send a request to an endpoint outside the EU. The categories this deployment uses. Inference: the endpoint that serves each request — a first-party endpoint in an EU region, or a partner endpoint operated outside the EU. Hosting: the provider whose servers run this site, the router and the database. Mail: the SMTP transport used for account messages such as verification, password recovery and invitations. Payments: none — card payments are disabled on this deployment, so no payment processor receives data. Error tracking: none installed; failures are written to this deployment's own logs. Analytics: none; this site loads no analytics script and sets no advertising cookie. The contracted legal identity, the processing location and the review date for a specific deployment are named in its data processing agreement, which is the document that binds, rather than inferred from this page.