DOI lookup
See the record a DOI is actually registered to, straight from CrossRef.
What resolving a DOI proves, and what it does not
A DOI has two halves. The prefix always begins 10. and names the organisation that registered it, so 10.1037 belongs to the American Psychological Association and 10.1038 to Springer Nature. The suffix is whatever the publisher picked, and publishers are advised to make suffixes opaque so nobody reads meaning into them.
That has one consequence worth carrying around: a DOI is real because somebody registered it, not because it is well formed. There is no checksum. Nothing about an invented DOI looks invalid, which is why a citation with a tidy identifier still tells you nothing until you resolve it. This page resolves it against CrossRef, the registry most journal publishers deposit with, and shows you the record that comes back.
Read the record, not just the fact that a record exists. The commonest quiet failure is a DOI that resolves perfectly to a paper nobody meant to cite, because the reference was built around a borrowed identifier or quotes the DOI of the whole issue rather than the article. If the title that comes back is not the paper in your citation, the citation is wrong, however well the link works.
When a DOI comes back with nothing
Clean the string first, since a fair share of failures are damage in transit. Strip a trailing full stop carried over from the end of a sentence, a doi: label from an older citation style, a hyphen a PDF inserted at a line break, a curly quote, a stray space. Then try again.
If it still comes back empty, there are a few honest possibilities. The paper may have been published within the last few weeks, before the registration propagated. The publisher may have moved the journal to a new platform and broken the registration, which happens more than you would think. Or the DOI was never registered because it was generated rather than looked up. The way to tell them apart is to search for the title: if the article exists and prints that exact DOI on its own page, the infrastructure is at fault, and the full decision tree runs through all five causes.
A dead DOI on a paper that does not exist anywhere is a different matter. That combination is the strongest fabrication signal available without reading the paper, and it rarely arrives alone, so check the rest of the list.
What this tool does not do
It reads CrossRef only, so a work registered with another agency, most datasets and some institutional repositories among them, can be genuine and still come back empty here. It says nothing about whether the paper is any good, whether the venue is a predatory or hijacked journal, or whether the paper has been retracted. And it cannot tell you whether the paper supports the sentence you cited it for. For a whole bibliography rather than one identifier, use the citation checker, which resolves every DOI in the list and searches for the references that do not have one.
Identifiers that came from the record
CiteOwl reads the publisher's metadata before it writes a reference, so the DOI in your bibliography was retrieved rather than predicted. Every change it makes arrives as a diff you can reject.
Start writing
