SeaLinkSeaLink
/
← Blog

2026-05-09 · SeaLink engineering · 5 min read

What clients see when a model is unavailable

Consistent errors, request IDs, and the product contract for reliable AI calls.

The reality

Model APIs can be temporarily unavailable for reasons outside a customer's control. SeaLink's public behavior is intentionally simple: consistent errors, request IDs for support, and status pages that show current availability.

Client contract

You keep using the same SeaLink base URL, API key, and model ID. The public contract is API behavior: OpenAI-compatible request and response shapes, consistent errors, and request IDs for support.

Handling unavailable models

When a model is unavailable, you receive a stable error shape and a request ID. Your app can retry, switch to another SeaLink model ID, or show its own service message. /docs/error-codes documents the failure behavior your app can handle.

Why we publish this

Two reasons. First, production teams need predictable behavior when a dependency is unavailable. Error format, request IDs, and status pages matter more than vague reliability claims. Second, some applications need stricter uptime terms or dedicated operational controls. SeaLink is the right choice when you want one API, clear error behavior, and billing records your finance team can reconcile.