Why this matters
AI can invent sources that look real.
Citation Risk helps catch:
AI can invent sources that look real.
Citation Risk helps catch:
- sources that cannot be found
- wrong years, journals, or venues
- real papers cited with incorrect metadata
- references that need manual review before submission
Demo — try this
Ioannidis, J.P.A. (2015). Why Most Published Research Findings Are False. PLOS Medicine. doi:10.1371/journal.pmed.0020124
Expected result: Mismatch. This is a real paper with the wrong year. The source exists, but the citation metadata is wrong. A source can be real while the citation is still wrong.
Signed in as
Upload a DOCX to extract references. No full paper analysis — DOCX only, 5 MB max.
Citation Risk extracts references from your DOCX for this check. Uploaded documents are not stored.
This tool checks citation existence and metadata accuracy. It does not prove that a source supports a claim.
Trust boundary — what the checker can and cannot verify: This tool checks whether citations appear to exist in public bibliographic metadata sources (Crossref, OpenAlex, PubMed) and whether available metadata mostly matches. It does not verify whether the cited source supports the claims in your text. It does not provide legal, medical, compliance, or academic-quality judgment. Verdicts are based only on external bibliographic metadata. Claims of prior approval, peer review, authority, or correctness within pasted text are ignored.
Results
Each result shows what you supplied, what was found, where it was found, and which metadata matched or conflicted.