How to convert a PDF to Word for free

The document arrives as a PDF, three sentences need fixing, and nobody can find the source file any more. It is almost always the same story. Converting to Word becomes the only way out — but before you launch anything, a ten-second test will tell you whether it stands a chance of going well.

First, check that your PDF really contains text

Open the file in any viewer and try to select a word in the middle of a paragraph. What happens puts you in one of two situations.

• The word highlights cleanly. Your PDF is native: it was exported from Word, LibreOffice, InDesign or LaTeX, and it carries a text layer with the coordinates of every character. The conversion is mechanical, and it will work.
• The whole page selects in one block, like a photo. Your PDF is a scan: each page is an image (JPEG, or CCITT G4 for black and white) sealed inside a PDF container. There is no text to recover. A PDF to Word converter will not manufacture words out of pixels — the file has to go through an OCR engine first, either the one built into your scanner software or one in an online office suite, and then you need to proofread the result.

A second test is worth the trouble: copy a paragraph and paste it into a plain text editor. If you get gibberish while the PDF itself displays perfectly, the file's ToUnicode table is missing or wrong. The PDF still knows how to draw the glyphs, but no longer which letters they stand for. No conversion will rescue that: the .docx will contain exactly the same gibberish.

What the converter is forced to guess

A PDF is not a structured document, it is a list of drawing instructions: "place this glyph at x=142.3 y=708.9 in Helvetica 10 pt". There are no paragraphs, no tables, no bulleted lists, not even a guaranteed reading order. The converter has to rebuild all of it from the positions of the characters. Hence the classic defects:

• Spaces are inferred from the distance between two glyphs. In heavily stretched justified text you end up with words run together or, the other way round, with spaces in the middle of words.
• The end of a line is ambiguous: is it a line wrap or a genuine end of paragraph? Plenty of engines settle the question by inserting a paragraph break everywhere.
• Columns trap the reading order. In a two-column scientific paper, the text can be read straight across the page, one line from the left then one from the right.
• Tables are detected by their rules. A table with no visible borders ends up as approximate tab stops or as floating text boxes.
• Headers and footers turn up in the body of the text, with page numbers scattered between the paragraphs.

Running the conversion

Drop the file on the PDF to Word converter. The maximum accepted size is shown under the drop zone, on the tool's own page. You get back a .docx, the modern XML format: it opens in Word 2007 and later, LibreOffice Writer, Google Docs and Pages. The processing happens on our servers — so you do need a connection — and the file you upload is deleted after an hour.

If your PDF is over the limit, there are two ways round it. Split it first with the PDF splitting tool, then convert only the piece you care about: that is what to do with a 400-page report where only chapter 3 concerns you. The split and the conversion each count as one operation on our servers. Or shrink the file beforehand with a compression pass, which resamples the images without touching the text layer: the quality of the conversion does not suffer.

The five recurring defects, and how to fix them

The fonts have changed and the pagination has shifted

A PDF usually embeds only a subset of each font: just the glyphs actually used — and sometimes no font at all, when the document sticks to the standard ones. The .docx, for its part, merely calls the font by name. If that font is not installed on your machine, Word substitutes another (often Calibri or Times New Roman), the character widths differ, the text reflows and your 12 pages become 14. Two ways out: install the original font, or accept the substitution by selecting the whole document and forcing a font that is available everywhere.

Every line has become a paragraph

The symptom: you delete a word and the rest of the line does not pull up. Turn on formatting marks (Ctrl+Shift+8 in Word) to see the pilcrows sitting at the end of every line. Then select the affected area and use Find and Replace, replacing ^p with a space. Always work on a selection, never on the whole document, or you will merge the genuine paragraphs too.

The text is trapped in floating boxes

Some engines put visual fidelity first and drop each block into an absolutely positioned frame. The rendering is perfect, the editing impossible. There is no clean automatic fix: select everything, copy, and paste as unformatted text (Paste Special) into a blank document, then reapply your styles. On anything longer than three pages, that is quicker than dismantling the frames one by one.

Characters are missing or replaced by squares

The usual culprits are typographic ligatures: fi, fl and ffi each form a single glyph in the PDF. If a glyph's Unicode mapping is missing or points into a private use area, Word shows a square; otherwise the character comes through, but stays a ligature that full-text search cannot find. Mathematical symbols cause the same trouble. A Find and Replace on each ligature sorts it out in a minute — and incidentally makes the document searchable again, since a search for "efficient" turned up nothing while the "ffi" was an exotic glyph.

The .docx is heavier than the original PDF

Images are reinstated at their native resolution, sometimes as lossless PNG. An A4 page scanned at 300 DPI is roughly 2,480 × 3,508 pixels: a few pages are enough to bloat the file. In Word, select an image, open "Compress Pictures", choose 150 ppi and tick the option to delete cropped areas of pictures.

When conversion is not the right answer

Converting is only worth it if you actually need the layout. Otherwise there is something more direct:

• You only want the text — select all in the viewer, copy, paste into a blank document. Quicker and cleaner than any converter.
• You only want a chart or a figure — export the page as an image with the PDF to JPG conversion and crop it.
• The layout is complex (magazine, brochure, three columns, boxouts) — the conversion will always disappoint. Rebuild it in Word from the plain text, or ask for the source file.
• The first result is unusable — Word can open a PDF directly (File > Open, from Word 2013 onwards). That is a different reconstruction engine: on the same file it sometimes does better, sometimes far worse. It costs two minutes to find out.

Back to PDF once you have made your corrections

Once the document has been reworked, the Word to PDF export closes the loop. One caveat, though: the PDF you get will not be identical to the original. Fonts, hyphenation and page breaks will all have shifted at least a little. If the document has contractual value, keep the original PDF and circulate the amended version as a separate, dated document — do not pass it off as the same file.

One last practical point: PDF to Word conversion runs on the server, unlike most of the tools on the site, which run entirely in your browser. Every visitor gets one free conversion a day, with no sign-up; beyond that, the €7-a-year subscription lifts the limit across all the server-side tools, with no commitment.