technique reference · by file type
Hidden text in RTF files
Rich Text Format files have their own ways to hide text: a control word that marks text as hidden, white text, tiny fonts, frames placed off the page, and sections a normal reader is told to skip.
Word and other editors honour those instructions and show nothing. Tools that extract the text often don't, so an AI model reads what the reader never saw. Below is every way we detect hidden text in RTF files.
rtf.document-propertyA disclosure that the file records a title, subject, author, company, the application that saved it, or a custom property — text a reader never sees in the body.rtf.hidden-vRTF has a control word whose entire job is "do not display this", and the text it hides is ordinary text in the file.rtf.ignored-destinationA `\*\…` section whose name the specification does not define — content a well-behaved reader is told to skip and a parser consumes.rtf.offpage-textA framed block placed by absolute coordinates that put it outside the paper the document declares.rtf.structural-destinationWords inside one of the ignorable sections the RTF and Word specifications define — a style table, a bookmark name, a footnote separator — reported quietly rather than as an alarm.rtf.tiny-fontA run set to a font size no one reads at, still carrying full words to anything that parses the file.rtf.unicode-fallbackText parked in the slot RTF reserves for readers that predate Unicode — discarded by every modern reader, emitted by naive extractors.rtf.white-textA run whose colour matches the ground it sits on — invisible on screen, ordinary text in the file.