Trust Center

Clear answers about how Modelflare handles your data.

Learn how Modelflare handles API data, protects credentials, uses service-provider categories, and manages retention, diagnostics, and security incidents.

How requests move

A request passes through authentication, routing, metering, and the model or infrastructure provider needed for the selected route. Request-level usage, status, latency, and cost metadata are recorded while standard request and response content is not retained as a long-term application database record.

At a glance

  • API request and response content

    In the standard API relay path, request and response content is processed transiently and is not stored as a long-term business record in the application database.

  • Havenbyte model training

    Havenbyte does not use Customer Content to train, fine-tune, distill, benchmark, or improve models developed or operated by Havenbyte.

  • Request diagnostics

    Short-lived diagnostics contain parameter summaries and diagnostic metadata, not complete request or response content, and are designed to expire after 72 hours.

  • Request Archive

    Request Archive is disabled by default. If specifically authorized and enabled, its current default expiry is 168 hours.

Security and provider boundaries

Modelflare uses authenticated access, protected credentials, encrypted transport, metadata-focused operational logging, controlled production procedures, and incident-response processes. Exact infrastructure topology, credential implementation details, detection thresholds, proprietary route mappings, supplier purchasing terms, backup configuration, and internal incident playbooks remain confidential.

Contact

Security, privacy, and enterprise trust questions may be sent to [email protected].