PDF to Word
What PDF to Word actually does
The reason an converted .docx is painful to edit is rarely the words. It is that each printed line arrived as its own paragraph, so adding a sentence pushes text nowhere and deleting one leaves a ragged hole; that the heading styles were never defined, so the navigation pane is empty and Styles shows nothing to click; and that a table came through as tab-positioned text, so it cannot be sorted, cannot gain a row, and falls apart the moment a column is widened. This page fixes all three before writing the file. It reads the document once, shows what it found as a list of blocks, and lets each wrong guess be corrected on the spot. What is exported is a Word document in the ordinary sense: Heading 1, 2 and 3 are real styles, bullets are a real numbered list, a detected grid is a real table with a header row, and the section is set to the size of paper the source was drawn on.
How to use it
- Choose a PDF. It is read in this tab, and what was found appears as a list of blocks with the page each came from.
- Correct anything read wrong. A block can be retyped as a heading or body text, its wording edited, a wrongly split paragraph merged, a wrongly joined one split, and a stray block dropped.
- Download the .docx. It opens in Word with a working navigation pane, editable tables, and paragraphs that reflow as they are typed into.
Useful for
- Pick up editing a report that only exists as a PDF, without first spending an hour deleting line breaks.
- Recover a price list or a specification table into a form that can be sorted and given new rows.
- Hand a colleague a document they can navigate by heading and mark up with tracked changes.
Limits worth knowing
- Headings are inferred from how much larger the text is than the body. A document that signals its structure only by colour or by indentation needs a few blocks retyped by hand.
- Merged cells, nested tables and columns that span are not rebuilt. A grid is recovered when several rows share the same column positions.
- Pictures, charts, footnotes and real Word header and footer parts are not carried across. A repeated running head is identified and left out rather than pasted into the body.
- A scanned page holds no text to read, so it yields nothing and says so rather than producing an empty file.
Questions people ask
Is the file uploaded?
No. It is read and the Word document is assembled inside this tab. Nothing leaves the browser.
Why is the navigation pane empty in files from other converters?
Because they apply a heading style without defining it in the document. Word then has nothing to build an outline from. The file this page writes defines the styles it applies.
Will the text still reflow when it is edited?
Yes, provided the wrapped lines were rejoined. That is on by default and can be switched off if the original line breaks matter more.
Which paper size does the result use?
The size the source pages were drawn at, so an A4 original stays A4. US Letter or A4 can be forced instead.