Crossref is a not-for-profit membership organisation that registers Digital Object Identifiers (DOIs) for the majority of scholarly literature and operates a set of infrastructure services built on top of that DOI metadata — reference linking, retraction alerts, plagiarism screening, and a public metadata API. It is not the DOI system itself, and confusing the two is the most common mistake researchers and administrators make when talking about it.
What Crossref Is, and What It Isn’t
The DOI system is a global identifier standard (ISO 26324) governed by the International DOI Foundation (IDF) and resolved through the shared Handle System. The IDF does not register individual DOIs itself — it licenses a small number of DOI Registration Agencies (RAs) to do that. Crossref is one of those RAs, and by far the largest by volume: it registers roughly 90% of active DOIs, covering journal articles, books, conference proceedings, and related scholarly content. DataCite is the other RA researchers encounter most often, covering research data, software, and other non-article outputs.
So “Crossref” and “DOI” are not interchangeable. A DOI can be registered by any of the world’s DOI RAs; a Crossref DOI specifically is one registered through Crossref’s infrastructure by one of its member publishers, and it comes bundled with the deposited bibliographic metadata Crossref requires as a condition of membership. That metadata — not just the identifier — is what makes Crossref useful to researchers, publishers, and research administrators alike.
Crossref vs DataCite
Both are non-profit DOI Registration Agencies operating under the same DOI Foundation/Handle System infrastructure, but they serve different deposit communities and metadata models. The comparison below is a quick summary — see the full Crossref vs DataCite comparison for a deeper side-by-side.
| Dimension | Crossref | DataCite |
|---|---|---|
| Primary deposit community | Journal publishers, university presses, societies | Data repositories, universities, data centres |
| Typical content registered | Journal articles, books, conference proceedings, preprints, datasets tied to an article | Research datasets, software, dissertations, other non-article outputs |
| Membership model | Publisher/member pays annual fee plus per-DOI charges | Member organisation (often via a consortium) pays annual fee plus per-DOI charges |
| Core services beyond registration | Cited-by reference linking, Crossmark, Similarity Check, Funder Registry | DataCite Commons (PID graph search), DOI Fabrica, event data |
| Metadata schema | Crossref XML deposit schema (currently 5.5) | DataCite Metadata Schema (XML, also served as JSON) |
In practice a single research output can touch both: a journal article gets a Crossref DOI, while the dataset behind it gets a DataCite DOI, and the two are linked through relatedIdentifiers/reference metadata deposited on either side. See the identifier crosswalk guide for how DOI, ORCID, and ROR identifiers connect across a research record.
How Publishers Use Crossref
Publishers join Crossref as members and register a DOI for each citable output at or shortly before publication. Registration means depositing structured metadata in Crossref’s XML deposit schema (currently version 5.5), which supports fields including title, author names, ORCID iDs, affiliations tagged with ROR IDs, the reference list, funder information (via the Crossref Funder Registry), licensing terms, abstracts, and — since schema 5.5 — native CRediT contributor-role tagging alongside Crossref’s own contributor-role vocabulary. Members are also expected to keep records current: correcting errors, registering retractions and corrections, and updating reference lists as they’re matched. The full mechanics of what to submit and when are covered in the Crossref metadata deposit workflow guide; the CRediT-specific markup is covered separately at Crossref Schema 5.5 & CRediT Contributor Roles.
Metadata quality is not cosmetic. Downstream services — citation counts, reference linking, discovery tools, institutional repositories, and CRIS platforms — are only as good as what a publisher actually deposited. A missing ORCID iD, an unmatched reference, or an absent funder ID propagates as a gap everywhere that record is reused.
How Researchers and Institutions Retrieve Crossref Metadata
Simple Text Query
For a one-off lookup, Crossref’s Simple Text Query tool lets anyone paste an unstructured reference or reference list — up to 1,000 references per submission — and get back a matching DOI for each, with no Crossref membership, account, or fee required. It’s the fastest way to resolve a messy bibliography into DOIs by hand. See the full Simple Text Query page for formatting tips and how ambiguous matches are handled.
The REST API
For anything programmatic, the free Crossref REST API (api.crossref.org) exposes the same metadata corpus Simple Text Query draws on, plus far more. The /works route is the core endpoint: query it by DOI, by field filters (funder, license, published date, container title, and more), or with the same free-text bibliographic matching Simple Text Query uses, exposed as the query.bibliographic parameter (e.g. api.crossref.org/works?query.bibliographic=<citation string>). No API key is required for the public tier; a higher-throughput “polite pool” is available by supplying a contact email, and a paid Metadata Plus tier exists for high-volume commercial use. This is the layer most reference managers, citation tools, and institutional systems integrate against directly rather than scraping publisher sites.
Why this matters for CRIS ingestion and institutional reporting
For research administrators, Crossref metadata is frequently the actual data source behind a current research information system (CRIS), not just a convenience lookup. Publication lists auto-populated into a CRIS or faculty activity system, funder-compliance reporting that needs to show a grant produced a given output, and institutional repository deposit workflows that de-duplicate against existing DOIs all depend on what a publisher deposited being complete and correctly matched to the right author (via ORCID), institution (via ROR), and funder (via the Funder Registry). A publisher that deposits thin metadata — no ORCID iDs, no funder data, an unlinked affiliation string instead of a ROR ID — creates work downstream: manual reconciliation, missed reporting, or a researcher’s output silently failing to appear where it should. This is the practical reason metadata completeness, not just DOI registration itself, is the thing research offices should actually be pushing publishers and their own repository/CRIS integrations toward.
Reference Linking and Cited-by
Because members deposit reference lists alongside their own metadata, Crossref can match a citing work’s references against other registered DOIs and build a citation graph across publishers — its Cited-by service. This is the infrastructure many citation counts and “cited by” links on publisher sites and aggregators are built from, and it depends entirely on reference-list deposit quality: an unmatched or malformed reference simply doesn’t link.
Crossmark: Version and Status Alerts
Crossmark is the Crossref service publishers use to signal a record’s current status directly on the published version — whether it has been corrected, retracted, or has an associated expression of concern — via a button linking back to the authoritative version-of-record information. In September 2023 Crossref acquired the Retraction Watch database, combining it with Crossmark-reported retraction data into what the two organisations describe as the largest single open-source database of retractions; retraction and correction records from that combined dataset are now available through the Crossref REST API. For a research office, this means retraction status can be checked programmatically against a DOI rather than relying on a publisher notice being seen manually.
Similarity Check
Similarity Check is Crossref’s plagiarism-screening service for members, built on iThenticate’s text-comparison engine and run against the Crossref-member full-text corpus and web sources. It’s a service layered on top of Crossref’s identifier infrastructure rather than a metadata-retrieval tool, and it’s frequently confused with commercial classroom plagiarism tools. See Crossref Similarity Check vs Turnitin for how the two actually differ in scope and audience.
Is Crossref Free to Use?
Retrieval is free: Simple Text Query, the public REST API tier, and the metadata itself carry no charge or account requirement for a researcher, librarian, or administrator looking things up. Registration is not: publishers pay Crossref an annual membership fee plus a per-DOI charge to register and maintain content, which is how the organisation funds the shared infrastructure everyone else uses for free.
Frequently Asked Questions
Is Crossref the same thing as a DOI?
No. A DOI is a persistent identifier standard; Crossref is one of several organisations (a DOI Registration Agency) authorised to register DOIs, and it happens to register the large majority of them for scholarly journal content. A DOI can also be registered through DataCite or one of the other DOI RAs, in which case it is not a “Crossref DOI” even though it is still a valid DOI.
How do I find a DOI for a reference without a citation manager?
Paste the reference into Crossref’s Simple Text Query form, or, if you’re doing this programmatically, call the REST API’s /works endpoint with the query.bibliographic parameter.
What’s the difference between Crossref and DataCite?
Both are non-profit DOI Registration Agencies, but Crossref’s deposit community is overwhelmingly journal/book/conference-proceeding publishers, while DataCite’s is overwhelmingly data repositories and universities registering datasets, software, and similar non-article outputs. See the full Crossref vs DataCite comparison.
Does Crossref detect plagiarism?
Only through Similarity Check, a separate member service built on iThenticate, not through DOI registration itself.
How does Crossref metadata affect a CRIS or institutional repository?
Most CRIS and repository ingestion workflows pull directly from Crossref (and DataCite) metadata to populate publication records, match outputs to authors via ORCID, and link outputs to funders and grants. Gaps in what a publisher deposited — missing ORCID iDs, missing funder IDs, unmatched references — surface downstream as incomplete or unlinked records in those systems.
Last verified 16 August 2026 against Crossref’s own documentation (crossref.org/documentation) and support articles (support.crossref.org).
