PDF to Word

A PDF does not understand the concept of a "paragraph" or a "sentence." A PDF is just a list of absolute visual instructions telling a printer where to drop ink.

For example, it simply says: "Draw the letter H at coordinate (100, 500), then draw e at (110, 500)". The fact that those letters form the word "Hello" in a paragraph is completely unknown to the PDF structure itself.

To convert this layout back into a fluid Word document (OOXML format), our extraction engine has to perform complex spatial heuristics.

It looks at the exact coordinates of every single character on the page and guesses which ones are close enough together horizontally to form a "word." It then looks at the vertical spacing between words to guess what constitutes a "line," and finally groups lines to form a "paragraph." This reverse-engineering is why PDF-to-Word conversions are rarely 100% perfect, especially with complex multi-column layouts.

If you scan a piece of physical paper, the resulting PDF does not contain text characters at all. It only contains a single large photograph (an Image XObject) of the paper.

Because this tool runs entirely offline in your browser for absolute privacy, it does not include a heavy Optical Character Recognition (OCR) machine learning engine. Therefore, if you convert a scanned PDF, the resulting Word document will simply contain that single large image, rather than editable text.