PDF to Word

No. It puts the text back in reading order using Word’s own styles, on a page the same size as the original with margins taken from where the text actually sits. Line breaks, exact fonts, letter spacing, columns and anything drawn as vector art are not reproduced.

They do not come out as Word tables. When a line has wide gaps inside it, those gaps become tab characters and the line becomes its own paragraph, so the data is all there and in the right order. To get a real table back, select those lines in Word and use Insert → Table → Convert Text to Table with Tabs as the separator. The report says how many lines came out that way.

Each column is read top to bottom in turn, and it is reported so you can check. Three columns work the same way. What breaks it is a page where the columns start and stop halfway down, or a sidebar that overlaps the main text; those come out interleaved.

A scanned page has no characters in it to read. Tick “Read scanned pages with OCR”, say which language your scan is in, and those pages get recognised instead. It reads one language at a time, so the wrong pick comes back as nonsense. It takes several seconds a page, and what comes back is plain text with no headings and no bold.

They come from the font each run of characters was drawn with, so they are right whenever the document used a real bold or italic font. They are wrong when a PDF faked bold by drawing each outline twice, and they can be wrong for a slanted regular font. Underline, text colour and highlighting are not carried over.

Yes. Each PDF becomes its own .docx. Two or more arrive as a single zip. A PDF over 2 GB cannot be opened at all; that is the browser’s limit.

No. The PDF is read off your disk by this page, the Word file is assembled in the tab, and the download comes out of your own browser. You can watch the network tab while it runs. The PDF reader and the OCR engine are downloaded from this site the first time; your document is not part of that.