Some documents are built to fool a plagiarism checker: the page looks normal, but the text a checker reads is not the text you see. Plagiarism Detector reads both - and shows you exactly where they differ.
A university evaluator sent us a thesis that passed his checks with a suspiciously clean report. Inside: 31 hidden pieces - quote marks in 1-point white text that turned 43 copied words into a "quotation", and 1-point "(2025)" glued after sentences to break every match. Today Plagiarism Detector finds all of them, removes them before the check, and lights the warning in the report.
The words are in the file, but nobody can see them on the page.
Text set to 1 point is a dot on the page but a whole word for a checker. Glued between two sentences, a 1-point "(2025)" breaks every match across it - and a 1-point quote mark can turn plain text into a fake "quotation".
White letters on a white page are invisible - and so are near-white ones like #F7F7F7, which a simple "is it white?" test misses. Inserted inside a word, they split it for the checker, so the word no longer matches its source.
Yellow letters on a yellow highlight read as a plain yellow bar.
Blue letters on blue shading - or any colour on the same colour - cannot be read.
A grey paragraph with grey letters inside it, or a page colour with letters in the same colour.
A character scale of 1 %, or a spacing of minus 10 points, stacks the letters into a hairline.
A 1-point space inside "plagiarism": you read one word, a checker reads two - and matches neither.
Converters and many checkers drop equation objects, so a whole paragraph typed as an "equation" disappears from the check.
A .docx should never carry macros. One that does can show different text when it is opened than the text that was checked.
PDF draws letters one by one - and can draw them so that nobody sees them.
Letters drawn in the colour of whatever is under them - the white page or a coloured box.
PDF can draw a letter with neither fill nor outline. It is used honestly for the text layer of scanned pages - so Plagiarism Detector only flags it when there is no scan underneath.
Words placed beyond the page edge are never shown or printed - but they are still in the file.
The PDF twin of trick 1: letters of 1-2 points.
Letters that look normal, or are not there at all, but change what a checker reads.
Cyrillic а, е, о, р, с look exactly like Latin a, e, o, p, c - but they are different characters, so "plagiаrism" with one Cyrillic letter matches no source. Plagiarism Detector repairs the word and shows the letter with its code.
Zero-width spaces and joiners have no width at all. One inside every word, and no word matches its source.
Combining grapheme joiners, variation selectors and Unicode "tag" characters (which can even carry hidden text) - all invisible inside a word.
Left-to-right marks, hair spaces so thin the word looks whole, a soft hyphen in every word - each one splits words for a checker.
Fullwidth letters and mathematical bold 𝐚 look like normal ones but are other characters.
Every trick is found where the text is read: in the Word file's formatting, in the PDF's letters, in the characters themselves.
The hidden text is taken out and look-alike letters are repaired, so the check compares the text you actually see.
In the Originality Report every spot gets a violet label: what was hidden, and how.
The report header lights "Counter-detection suspected!" with the count - you cannot miss it.
Not Plagiarism Detector. White, near-white and background-coloured text is found from the document's own formatting, removed before the check and marked in the report.
No. Plagiarism Detector checks every word for letters from another alphabet, repairs the word, and shows each swapped letter with its Unicode code.
Yes - zero-width spaces and joiners, direction marks, hair spaces, soft-hyphen floods, CGJ, variation selectors and Unicode tag characters inside words.
The rules were tuned on real documents to leave formulas, scientific notation, web links, scanned pages and university cover pages alone. A few stray invisible characters from copy-paste are cleaned silently, not flagged.
In the Originality Report: a violet label at every spot in the text, the "Counter-detection suspected!" warning in the header, and a list by type in the UACE section.
Yes. Download our demo Word file - every trick on this page is inside it - and check it with Plagiarism Detector.
Plagiarism Detector for Windows 3087 and later.
How students try it, and why it fails: How Students Cheat · How Detection Works · Research & Track Record