Audit one research paper.

Start with the publisher or repository page

Use the page for one article, preprint or other research paper. The free scan reads one public HTML page, subject to access rules and request limits. A publisher’s article listing or a researcher’s profile can describe several works, so choose the page that identifies the specific paper.

Add up to five questions, such as “Which dataset was used?” or “What limitation did the authors report?” A page containing only the abstract may provide enough evidence for some questions. The report keeps the source passage so you can judge what it actually answers.

Inspect metadata with its source

The report records supported title, author, publication date, DOI, abstract, canonical URL and PDF-link metadata from the HTML. Each extracted value keeps its source label. Multiple values remain visible so you can spot a preprint DOI mixed with a later publication, or a title inherited from a different page.

Metadata helps systems identify a scholarly work. Google Scholar publishes separate technical guidance for article pages and their bibliographic metadata. A CiteMaybe scan does not query Scholar’s index. Google Scholar inclusion guidelines.

Check the identity behind a citation

Paste up to three DOI identifiers or public HTTPS reference links into the private report. The check compares those inputs with the page-declared paper identity, without querying a DOI registry, and reports a matching DOI, an exact page match, a different DOI or an unverified result. Another article on the same publisher’s domain does not establish a match.

A DOI identifies an object persistently. Separate versions can have separate identifiers, so review the version of the work you intend to cite. DOI Foundation’s explanation.

When checking an AI answer, keep the exact question, product, date, answer excerpt and source links in the free platform journal. You run the question in the product yourself. CiteMaybe checks the identity of the recorded references and preserves your observation.

Use the result within its scope

A metadata match cannot establish that an AI system cited, read or used the paper. This audit does not measure scholarly citation counts, h-index, peer-review quality or the accuracy of the research. Missing or conflicting identifiers remain unresolved until you review the source.

The scan checks server-delivered HTML. Access restrictions, an unavailable page or content supplied only through JavaScript can limit the evidence. Linked PDF files are recorded as links and are not opened. Your private report lasts seven days in the browser session that created it. Save the PDF or JSON export for a lasting copy.

Read the audit methodology · Private report access · Website audit pricing