LogoAI SDK Token Usage

Error handling

Error handling in AI SDK Token Usage follows the same predictable pattern as the data-fetching layer used internally — SWR. Just like the useTokenCost hook, which returns { data, isLoading, error }, every asynchronous operation in the library exposes its own error state in a structured and non-disruptive way.

Instead of throwing exceptions, errors are wrapped in a standardized TokenUsageError object that includes useful context such as the error name, message, status code, and additional info. This makes it easy to identify the source and type of a problem — whether it’s invalid metadata, missing pricing information, or a failed network request.

This page focuses on the different types of errors you might encounter when using the library. For practical examples of how to display or recover from these errors in a React app, see the demo application.

Error types

ErrorDescriptionTypical cause
FetchErrorAn error occurred while fetching data.Network request failed, timed out, or returned a non-2xx status.
ModelNotFoundErrorThe requested model could not be found in the model catalog.The provided model slug or ID is unknown or outdated.
MissingMetadataErrorMessage metadata is missing or invalid.The message lacks required TokenUsageMetadata fields (totalUsage, canonicalSlug).
CostComputationErrorCost computation failed due to missing pricing details.One or more models do not have pricing information in the catalog.
UnknownErrorA catch-all for any unexpected or unclassified error.An unhandled or internal logic failure.

If the experienced error is ModelNotFoundError or CostComputationError, take a look at the info field in the error object for more details, and inspect the model catalog at models.dev to verify that the model exists and has defined pricing information.

Error object shape

All errors in hooks share the same structure and extend the TokenUsageError type.
This ensures that you can reliably access common fields such as the error name, message, and status code, no matter which specific error occurred.

Prop

Type