Q&A evaluation dataset for the Enscrive developer documentation.
Queries represent typical questions a developer integrating with Enscrive's /v1 API would type. Each query is paired with the expected source document (a markdown filename) and the H2/H3 section anchor inside that document that most directly answers it.
Populated — v1.0.0, 2026-05-13. 44 query / expected-document pairs across 11 source markdown files. Croissant 1.0 descriptor… See the full description on the dataset page:
https://huggingface.co/datasets/Enscrive-IO/docs-eval-v1.
