eval

Creates accept an optional id: an identical replay returns the resource; conflicting id or identifier returns 409. Lists are paginated.

post/api/v1/eval/datasets

Request

  • Base URL: https://app.lobehub.com
  • URL: https://app.lobehub.com/api/v1/eval/datasets
  • Auth: HTTP bearer

Request body

idstring
identifierstring required
namestring required
benchmarkIdstring nullable
descriptionstring nullable
metadataobject nullable
evalConfigobject nullable
evalMode'equals' | 'contains' | 'regex' | 'starts-with' | 'ends-with' | 'any-of' | 'numeric' | 'extract-match' | 'json-schema' | 'javascript' | 'python' | 'llm-rubric' | 'factuality' | 'answer-relevance' | 'similar' | 'levenshtein' | 'rubric' | 'external' nullable

Response

Successful response

messagestring
successtrue required
timestampstring date-time required

Changes