PDF to Word

Turn a PDF back into an editable document — real paragraphs, headings and lists, not a page of text boxes. Opens in Word, Google Docs, Pages or LibreOffice. Your file never leaves your device.

The choice most converters make quietly, and get wrong A PDF can be reproduced in Word almost exactly by wrapping every line in an absolutely positioned text box. It looks perfect and it is unusable: change one sentence and nothing reflows, because nothing was ever a paragraph. This does the other thing — it infers real paragraphs and lets Word lay them out again. The result looks less like the original and is the only version you can actually edit. If you wanted the appearance preserved exactly, you wanted the PDF.

Your file

What it can and cannot work out

A PDF holds glyphs at coordinates. It has no paragraphs, no headings, no lists and no reading order. Everything here is inferred from where the text sits and how big it is.

Headings

Found by size relative to the body text, never by absolute points — a document set in 9pt and one set in 14pt have different bodies and the same structure. Bold and short also counts. Bold alone deliberately does not: plenty of documents bold an entire paragraph for emphasis, and promoting that to a heading wrecks the outline and the navigation pane built from it.

Paragraphs

A new paragraph starts where the line spacing grows, where the previous line stopped well short of the right margin, or where the kind of line changes. None of those is reliable alone; together they are usually right.

Words broken across lines

A line ending in a hyphen is usually a broken word, and sometimes a real hyphen. “informa-” and “tion” is one word; “well-” and “Known” is two. Lower-case on both sides is treated as a break; anything capitalised, numeric or punctuated keeps its hyphen. That rule is right far more often than it is wrong, and when it is wrong it is visible.

Columns

This is the failure worth understanding, because it does not look like a failure. A PDF stores text in drawing order, and on a two-column page the two columns often share their vertical positions. Read naively, you get alternate clauses from two different arguments spliced into sentences that parse perfectly and mean nothing — fluent nonsense rather than obvious rubbish.

Columns are found from the gap the text never crosses, and split before the lines are formed. Doing it the other way round merges the two columns into single lines and no reordering afterwards can undo it.

Images

Embedded pictures are extracted and placed among the text by where they sat on the page, at the size they occupied there rather than at their pixel size — so a 2000-pixel logo dropped into a 40-point box stays small.

Images under about 12 points in either direction are left out. A PDF is full of one-pixel images used as spacers and rules, and a document that faithfully reproduces every one of them is worse than one that quietly drops a logo.

Bold and italic

Read from the font’s name, which is all a PDF reliably offers. A document whose fonts are embedded as g_d0_f1 gives nothing away, and emphasis is lost on those pages. The tool says so when it happens rather than letting you find out later.

What it will not do

Scans. A scanned page is an image of text, not text — there are no glyph positions because there are no glyphs. The tool will tell you it found nothing rather than handing back an empty document. The quick test: try selecting a line in your PDF viewer. If you cannot, there is nothing here to convert, and the job you need is optical character recognition.

Vector artwork. Photographs, logos and other embedded pictures are carried across at the size they occupied on the page. A chart drawn with lines and curves is not a picture in the file at all — it is a set of drawing instructions — so there is nothing to extract, and it will not appear.

Tables. They come through as text rather than as Word tables. A table in a PDF is not a table either; for one specifically, the PDF table extractor does a much better job and gives you a spreadsheet.

Exact appearance. By design, as above.

Nothing is uploaded

The PDF is read by pdf.js and the Word file is written by the docx library, both running in your browser. Your document is never uploaded. There is no server here to upload it to.

To be exact about the requests that are made: the first time you use this page it downloads those two libraries from a public code CDN. Both happen before you choose a file, carry nothing about you, and would look identical whoever made them. After that the page runs entirely on your machine.