How to tell a fabricated DOI from a real one

The single fastest check on any reference that has one.

A DOI is a permanent identifier assigned to a work by a registration agency when it is published. It looks like 10.1086/225469: a prefix beginning 10. that identifies the registrant, a slash, and a suffix the publisher chooses. Because it is assigned centrally and never derived from the work, a DOI cannot be inferred, guessed, or reconstructed. Either it was registered or it was not.

That property makes it the strongest test available. A language model producing a citation can imitate a title, an author list, and a volume number, because those follow patterns. It cannot produce a registered identifier for a work nobody registered.

How to resolve one

Paste the DOI after https://doi.org/ and open it. Three outcomes:

It lands on the publisher's page for that exact work. The reference exists. You still want to confirm the title and authors match what you wrote.
"DOI Not Found." No registration agency has this identifier. On a citation that otherwise looks polished, this is the fabrication pattern.
It lands on a different paper. The citation was assembled from parts, or a digit was mistyped. Compare the resolved title against the one you have before assuming either.

Two traps

A well-formed DOI is not a real DOI. 10.1016/j.jbusres.2021.99999 has the correct shape and a real publisher prefix, and resolves to nothing. Shape tells you the string is syntactically valid, which is a different and much weaker claim.

A resolving DOI does not confirm the rest of the citation. The identifier points at one specific work. If the year, volume, or author list in your reference disagrees with the page it opens, the reference is still wrong, and a reviewer will find that as easily as you just did.

When there is no DOI

Books, older articles, working papers, and conference proceedings often have none, and their absence proves nothing. Fall back to the title, author, and container checks.

The shortcut: paste your bibliography into CiteFact and every entry is checked against CrossRef and OpenAlex at once, retractions included. Free for 10 references, no signup.

Method and coverage

Verdicts come from two sources and nothing else: CrossRef, the DOI registration agency for most scholarly publishers, and OpenAlex, an open index of scholarly works. Retraction status comes from the Retraction Watch database, which CrossRef has distributed since 2023. We run no language model anywhere in the checking path. One reference checked twice returns one verdict, and every verdict traces to a record you can open yourself. A reference we cannot find is reported as unverifiable when it belongs to a category those registries do not index, and as not found only when it is the kind of work they would normally index. The checker applies this method to a whole bibliography at once.

What your library can do that this cannot

This page and the checker behind it are a self-service first pass. They query two open registries, which leaves real gaps: consultancy and agency reports, non-English scholarship, dissertations, conference proceedings outside the major series, and anything published before those registries have coverage. A subject librarian and your institution's licensed databases index all of it, and can confirm the items we can only mark "unverifiable." If a reference matters to your argument and this page cannot settle it, that is the next stop.

Citing this page

CiteFact. "How to tell a fabricated DOI from a real one." citefact.com/fabricated-doi. Reviewed September 2026.
Permanent link https://citefact.com/fabricated-doi
Corrections and additions to contact@citefact.com

Librarians and instructors are welcome to link to or adapt this page for course and research guides. Tell us if something here is wrong or out of date and we will fix it.