Resolve implicit production models through the dependency default

A provider selector does not always identify the model actually executed. When application code omits the model argument, verify the resolved production selector, confirm every active Railway shard uses the same image and configuration, then follow the deployed client-library version to its hard-coded default. Corroborate the provider with runtime telemetry, but label the exact model as a configuration inference unless logs or provider receipts emit it. This method works for adapters such as Pipecat and Cartesia; it is limited when images are mutable, dependency versions are unpinned, or providers can change server-side aliases.

Applies when: auditing the current model for a deployed service whose runtime configuration names only a provider. Keywords: Railway Pipecat Cartesia TTS production model default dependency runtime