How to keep formatting when you convert PDF to Word

What survives the translation

A PDF stores the finished page: glyphs at coordinates. Word stores instructions: this run is a paragraph, this is a heading, these cells are a table. Conversion translates the drawing back into instructions, and the translation succeeds for the parts of the page that were simple to begin with.

In practice, straight body paragraphs, ordinary lists, and headings that were real styles in the original document come through most often. Fonts arrive approximately — the name may survive even when the exact cut does not. Page breaks land roughly where they were. That part of the file you can usually edit right away.

What usually does not

Multi-column layouts, sidebars, text wrapped around images, running headers and footers, annotations, and tables with merged or nested cells are where conversions lose their footing. The converter can see where everything sits on the page; it cannot always see what belongs together.

A scanned page is a separate case: there is no text to translate at all. The output is an image of the page, or a guess at the characters. This site treats scans as two jobs — make the PDF searchable first, review that text, then convert — and does not claim character-perfect recognition.

Why text boxes appear

When the converter cannot infer structure, its safest reconstruction is positional: put each run of text where the PDF put it. In Word, that means text boxes. The page looks correct in print preview because every line sits exactly where it did. The cost is editability — text in a box is not part of the document body, so typing into one paragraph does not move the next.

A few boxes are normal, usually around figures and labels. A box on every line of a letter means the converter handed back a photocopy. The longer guide on this site, “PDF to Word without text boxes,” shows how to tell the difference in under a minute.

How to inspect the file you got

Click into the middle of a paragraph. A frame around the line means that run is floating, not flowing. Open the navigation pane: if no headings appear, the outline was not recovered. Click a table and try to add a row; if the editor refuses, the grid was drawn, not built. Zoom in on anything that looks soft — pixelated letters are an image, not text.

Run these checks before you start editing, not after. Discovering at line forty that the whole section lives inside one box means redoing forty lines.

What Papermeld does with formatting

Papermeld converts the file in a conversation and puts a structure comparison beside the output — headings, tables, page order, and notes on what deserves a check. If the Word file has text only in boxes, or no real body text, Papermeld does not treat that output as a completed conversion.

Sign in from this page, upload the PDF, and say which parts of the formatting matter. If you already know the failure — “the columns collapsed last time,” “page three is a scan” — name it in the same conversation so it gets checked first.

Formatting you can repair afterwards

Some drift is cheaper to fix in Word than to prevent in the conversion. If headings arrived as bold body text, restyle them as real heading styles so the navigation pane and any table of contents work again. If the source font did not survive, pick one deliberate replacement and set it through styles rather than line by line. Ten minutes of structure repair turns a passable conversion into a file you can maintain.

What no converter can promise

No tool keeps every visual detail of a complex page and still produces an editable document. Position and editability trade against each other by design. The honest goal is a file whose structure you can verify — headings, paragraphs, tables — and whose remaining drift you can see and fix deliberately. You remain responsible for reading the output against the source before you send it anywhere.