Fake citation checker
Catch fabricated and non-existent references before they reach a reviewer, instructor or court. Paste your list — CiteProve flags every citation it cannot locate in an authoritative database.
In short
A fake citation is one that no authoritative database can locate. To detect them, verify each reference against Crossref, OpenAlex, PubMed and CourtListener instead of judging it by how well-formed it looks — a fabricated citation often has a flawless format and even a real-looking DOI. CiteProve checks a whole reference list at once and tells you exactly which sources could not be found.
Paste your reference list
The four kinds of fake citation
- Pure inventions — plausible authors and a journal that simply do not exist.
- Wrong-identifier fakes — a genuine DOI or PMID bolted onto a fabricated title.
- Mangled real papers — a real source with the wrong year, swapped authors or a near-miss title.
- Chimeras — real authors from one paper, the title of another, the journal of a third.
We flag — we never accuse
Calling a real source fake is the worst mistake a checker can make, so CiteProve is built to avoid it. Every verdict shows which databases were consulted with outbound links, two near-equal matches are reported as ambiguous rather than guessed, and anything outside our coverage is marked can't-check — neither confirmed nor flagged.
Frequently asked questions
- How do I know if a citation is fake?
- A fake citation cannot be located in any authoritative database. CiteProve resolves any identifier and searches Crossref, OpenAlex, PubMed and CourtListener; if no record matches the title, authors and venue, the reference is flagged as not found — carefully worded, because absence of a record is a flag, never an accusation.
- Can a fake citation have a real DOI?
- Yes. A common failure mode is a genuine DOI or PMID attached to a fabricated title. CiteProve resolves the identifier and then validates the other fields against the record it points to, so a DOI that resolves to a different work is flagged rather than trusted.
- Will it ever wrongly call a real source fake?
- We design against that. Our highest-precision target is the not-found verdict, because a false accusation is the worst error. Uncertainty always degrades to a softer verdict, and a book or non-English source we can't index is reported as can't-check, never as fake.
Related: ChatGPT citation checker · Check for retracted papers · Citation glossary