Crossref

Crossref is the organisation most publishers use to register a DOI (Digital Object Identifier), the permanent link that always points to a given article or book even if its web address changes. Registering a work also means depositing its metadata, and Crossref makes that metadata openly available for over 150 million scholarly works: journal articles, books, conference proceedings, datasets, and preprints.

Each record can include titles, authors, affiliations, funding sources, licences, abstracts, and full reference lists, which together map how the scholarly record connects. You reach it through a free REST API that takes rich queries and filters, or through the full metadata dumps if you would rather work offline. A courtesy tip: pass your email in the API request (the "polite pool") and you get faster, more reliable service. The metadata is released under CC0, placing it in the public domain, so you can use and redistribute it commercially with no attribution obligation, a genuinely rare freedom for data at this scale.

For RAG (Retrieval-Augmented Generation), the key thing to understand is that Crossref gives you metadata, not full text. That makes it less a source you retrieve answers from and more the scaffolding around a scholarly corpus: resolving citations to real works, deduplicating documents by DOI so you do not index the same paper twice, and enriching each chunk with reliable bibliographic detail your model can cite. Abstracts are present for some records but far from all, so do not rely on them being there.

Think of Crossref as the registry layer. Pair it with CORE or Unpaywall when you need the actual full text, or with OpenAlex when you want the same bibliographic picture plus a richer citation graph and author disambiguation already stitched together.