provider-availability-scout

You can monitor multi-provider model availability by scheduling a scouting process that checks vendors like OpenAI, Anthropic, Google, and OpenRouter for pricing shifts, deprecations, and new restrictions to keep routing accurate.

Diff-driven routing compares current provider observations to prior state to highlight model additions, removals, and policy changes. You need it to minimize silent degradation in a multi-provider inference runtime.

Yes, you can detect API deprecations and pricing changes across OpenRouter, Anthropic, and other listed providers by running a scheduled scout that surfaces diffs in model availability and restrictions.

You update routing defaults by running the scout weekly to refresh observations, generating diffs that highlight availability and pricing shifts, and applying those optional routing updates to your configured provider list.

Yes, you need a small state store to retain prior provider observations. This allows the system to compare current availability against historical data and generate accurate diffs for routing decisions.

Limitations include requiring a pre-configured provider list to function and relying on a small state store for diff generation, meaning it cannot track unlisted vendors or retain historical context indefinitely without storage management.