Model routes

A route connects a model with its execution and data handling path. Confirm the models and routes available to your integration with the Mystro team; this guide does not publish a universal model catalog or imply that every model supports private inference.

Mystro private inference routes use zero-retention model processing for inference-provider prompts and responses, including fallbacks. TEE and E2EE coverage remains specific to supported routes.

Confirm the route

For each workflow, record the approved model, TEE coverage, E2EE endpoints, request limits, tool support, and expected failure behavior. Verification and attestation details should be supplied for the actual supported route.

Handle unavailability deliberately

Design your product to show a useful unavailable state when its required route cannot run. Agree on any fallback explicitly and confirm its TEE and E2EE coverage before enabling it. Zero-retention model processing also applies to fallback routes.

Use non-sensitive examples to validate quality and failure behavior. Reconfirm route properties when changing models or enabling additional tools. Start with developer access.