Using the Inference API

How the inference API behaves: calls, request IDs, errors, limits.