Anonymity Breach Scanner: Check Blinded Manuscripts Free & Online

Double-blind review fails when the manuscript itself gives the authors away. This free heuristic scanner checks your excerpt for self-citation phrases, author names, institution mentions, and grant numbers before you submit.

What the tool does

Double-blind review only works if the manuscript itself keeps quiet. Authors routinely leak their identity in a phrase they no longer notice: a reference to their own earlier study, a mention of their lab’s dataset, or a funding line with a grant number a reviewer can search. This scanner reads an excerpt you paste and flags those leaks with simple, visible rules. It looks for self-citation phrases such as our previous work, for any author surname you list, for any institution name you list, and for grant-number patterns. Each hit becomes a row in a table with its type and the exact text that triggered it, so you can fix the sentence instead of guessing where the leak is.

How to use it

Paste the riskiest parts of the manuscript, not the whole paper at once: the introduction, the methods, and the acknowledgements are where leaks hide. Type the author surnames and institution names in the two small boxes, separated by commas. Press the button and work through the table top to bottom. Rewrite each flagged sentence in neutral language, for example replacing our previous work with prior work found, and rescan the edited text. Keep scanning until the table is empty, then do one final read by eye, because no word list catches everything.

Worked example

Suppose an excerpt says: Building on our previous work at Stanford University, we extend the dataset collected under grant NSF123456. The scanner flags three things at once: the self-citation phrase our previous work, the institution mention Stanford University, and the grant-number pattern. Fixing it is straightforward: Building on prior work, this study extends a dataset collected with public funding. The rescan returns zero flags. That three-minute edit can be the difference between a genuinely blind review and one where the reviewer knows exactly whose paper they are reading.

Reading the results honestly

Zero flags does not mean zero risk. A reviewer may recognise a distinctive method, a unique field site, or a dataset only one group could own, and no word matcher will catch that. Conversely, a flag is not proof of a leak: a surname may appear as a cited author in a normal reference, and an institution may be named as a study site. Read each flag in context and ask one question: would this sentence help a knowledgeable reader name the authors? If the answer is yes, rewrite it. If the answer is no, leave it.

Honest limits

This tool is a heuristic word matcher, not an anonymity guarantee. It only checks the text you paste and the names you provide, so a misspelt surname or a paraphrased self-reference will slip through. It does not read figures, supplementary files, file metadata, or author names hidden in document properties, all of which must be cleaned separately. Treat it as a pre-submission screen, and pair it with a careful manual read before you upload.

Practical tips for blinded manuscripts

Scan in stages: introduction first, then methods, then acknowledgements, because each section leaks in a different way. Introductions leak through self-citation, methods leak through unique equipment or field sites, and acknowledgements leak through funders and named helpers. Strip document metadata too: author names hide in file properties, tracked changes, and comments, and no text scanner can remove them. Ask a colleague who does not know the paper to read the blinded version and guess the authors; if they guess correctly, the leak is real even if the scanner found nothing. Finally, check your figures and supplementary files with the same care, since a lab logo on a photo or a familiar dataset name in a file can undo an otherwise clean manuscript.

Embed this tool on your website — free

Copy the code below to add this working tool to your own page. A small credit link stays with it — that’s all we ask.