Hugging Face hub

Hugging Face Dataset Rows Info API

Whether a dataset's rows can be shown, why they cannot, and the subsets, splits and columns behind them.

GET /v1/hugging-face/datasets/{owner}/{name}/viewer

1 credit Failed requests are free

About this endpoint

GET /v1/hugging-face/datasets/{owner}/{name}/viewer returns lookupStatus and viewer. viewer.status is available or unavailable, and viewer.reason says why when unavailable: disabled, no_data_files, empty, unsupported_format, too_large or failed.

When rows are available, viewer also carries isPartial, supports (preview, viewer, search, filter, statistics), totalRows, parquetBytes and configs. Each subset has name, numRows, numColumns, license, homepage, features and splits, and each split has name, numRows and numBytes.

Hugging Face Dataset Rows Info API FAQ

Why can rows not be shown for a dataset?

Read viewer.reason. The owner may have switched it off, the data files may be missing or empty, the format may be unsupported, the dataset may be too large, or preprocessing may have failed.

How do I know which subset and split to request?

Read viewer.configs. Each subset lists its splits; pass the subset as config and the split as split to the rows and statistics endpoints.

Is this the same as the dataset detail endpoint?

No. The detail endpoint returns the dataset record. This endpoint returns whether and how its rows are available.

How much does a row facts lookup cost?

Pricing is documented on the operation in the API registry (shown on this page). Confirm on every response with meta.creditsCharged: that field is the billing source of truth.