A .docx file is a ZIP archive of XML. Since Word 2003, Word has stamped a revision save identifier into that XML for each editing session that ends in a save — and it stamps it onto the paragraphs and runs that session touched.

That is why a document can be read backwards. If a paragraph carries the identifier 00A1B2C3 and the next twelve carry 00D4E5F6, those two stretches were written in two different sittings. A document built up over a month is a mosaic of identifiers. A document that arrived in one action is one flat colour.

What this page reports, and what it refuses to

There is a number in every .docx that looks like the answer and is not. The file's settings.xml declares a table of identifiers, and its length looks exactly like a count of editing sessions. It is inherited from whatever template produced the document.

We measured this rather than assuming it. A document produced by a generator, which no person ever edited, declares nine editing sessions in that table — and describes itself as written in Microsoft Word, because the template it was built from was. A real Word document in the same sample declares 288 while six are observable in its actual content.

So this tool ignores the declared table and reports only identifiers attached to your document's own paragraphs. Where the two disagree, it says so and explains why the larger number is not the answer.

What it cannot establish

What this cannot tell you

  • Word does not follow the ECMA specification for revision save identifiers, so the exact order of editing sessions cannot be recovered from the file.
  • A writing history shows that editing happened. It does not show who did the writing.
  • No writing history in a file is not a sign of wrongdoing. Exports, converters and generators all produce documents with none.
  • This is not a detection result. It does not answer whether any part of the document was written with AI assistance, and it cannot.