📃 PDF to Word Converter

About converting PDF to Word

A PDF has no paragraphs. It only has ink at coordinates, which is why so many converters hand back a document that looks right and cannot be edited. This one rebuilds the document instead: it reads the position, size, weight and colour of every piece of text, works out where lines become paragraphs, which of those paragraphs are headings, which are list items, and where the tables are, then writes real WordprocessingML. What you get back is a .docx full of genuine Word paragraphs and runs that you can click into and type over.

All of that happens in your browser. The PDF is never uploaded, so contracts, medical letters, bank statements and anything else you would rather not hand to a stranger stay on your own machine. You can watch it for yourself: open your browser tools, switch to the network tab, and convert a file. Your document never leaves the tab. The page itself does load Google Analytics, which counts the visit and never sees the file, its name or anything in it - and if you want a page with no analytics at all and a browser-enforced block on every outside request, our Zero-Egress tools do that.

What gets rebuilt

• Real paragraphs and runs: lines are regrouped into paragraphs using leading, indentation and alignment, and each run keeps its font, size, weight, slant and colour.
• Headings as heading styles: a heading arrives as Heading 1, 2 or 3, so the navigation pane and the table of contents work rather than just looking like big text.
• Lists as real numbering: bullets and numbered items are written into numbering.xml, so pressing Enter continues the list instead of typing another literal bullet character.
• Tables as real tables: column boundaries are found from the whitespace between cells and written out as a Word table with rows, cells and borders you can drag.
• Pictures: raster images are lifted straight out of the PDF, re-encoded as PNG and placed back at the size they occupied.
• Links, page size and orientation: link annotations become clickable Word hyperlinks, and the page size, margins and landscape or portrait orientation come across in the section properties.
• Reading order: a two-column paper is linearised properly, so column one is read to its end before column two begins.

Flowing or exact layout

Scanned PDFs

If a page is a photograph of paper there is no text to read, so the converter falls back to OCR that also runs on your device. Pages without a text layer are spotted automatically and listed in the report before you convert, and you can pick the recognition language. The engine downloads once and is then cached, which is why the first scanned conversion takes longer than the ones after it.

Be honest with yourself about OCR output: it is recognised text, not the original text. Numbers, names and anything set in an unusual face deserve a read through before you send the document on.

When a PDF cannot say what its own text is

A PDF does not store letters. It stores glyph numbers, plus a table saying what those numbers mean. Plenty of real documents ship with that table half filled in - Telugu, Hindi and Tamil land records, receipts and certificates are common offenders - and the usual result is a converter printing the glyph numbers as if they were letters. You get a Word file full of characters like ƻ, ŵ and ǐ that appear nowhere in the original. It reads as text, so nothing warns you.

This converter checks. Every glyph on a page is compared against what the PDF says that glyph means, and a page that cannot account for its own characters is reported to you before you convert. Where the PDF states the real wording separately, that wording is used and the result is exact. Where it does not, the affected lines are re-read from the page image with on-device OCR and flagged for proofreading. Where neither works the lines are left out. What never happens is a character being invented and handed to you as though it were in your document.

Frequently asked questions