ToAPIs Legal

Subprocessors and Data Regions

The current public boundary for model routing, upstream processing, and region information.

Last updated: July 23, 2026

1. Upstream processing

To provide the service, we may forward API requests to applicable upstream AI providers. The applicable provider depends on the selected model, availability, region, and runtime routing.

2. Routing is not a static commitment

We do not currently publish a static list that permanently binds a public model ID to one upstream provider, subprocessor, or processing region. Such a list could become inaccurate after capacity, failover, or compliance-routing changes.

We have not yet published a route-level statement for gpt-image-2, seedance-2-mini, and seedance-2-fast covering training, evaluation, or service-improvement use.

3. Enterprise diligence

For a specific model route, subprocessor, region, or upstream terms, contact [email protected] with the model, calling region, account, and expected data categories. We will provide applicable information after verifying the live production configuration.