PDF to DOCX Converter

pdf → docx

Turn a PDF into an editable Word (.docx) document. Your browser reads the text layer and writes the file itself, so nothing is uploaded.

Settings
QueueNothing is uploaded
Drop a PDF here to convert it
Reading PDF text layer…

About this PDF to DOCX converter

The converter turns a PDF into an editable Word document in your browser. PDF.js reads the text layer page by page, the positioned text is grouped back into lines and paragraphs, and the result is packaged as a real Office Open XML .docx file with JSZip.

A PDF records where each character sits on the page, not whether a line is a heading or a table row, so the converter has to infer structure from position and spacing. It gets reading order and paragraphs right on an ordinary text page. A scanned page has no text layer at all, so there is nothing for it to read.

How to use it

01

Add a PDF

Drop a PDF on the box, or choose it with Select PDF file. A text preview appears if the file has a text layer, or a warning if it doesn't.

02

Choose the layout

Pick whether wrapped lines merge into paragraphs or stay as separate lines, and whether to insert a "Page N" heading per page.

03

Convert

Click Convert PDF to DOCX. PDF.js reads each page and a progress bar tracks it.

04

Get your file

The finished .docx shows up with a Download button, named after the PDF you started from. It reaches your device only when you click.

What it does

  • Your browser reads the PDF and writes the Word file; nothing is uploaded.
  • Reading order and paragraph breaks carry over. Wrapped lines merge back into paragraphs by default, or you can keep every PDF line as its own paragraph.
  • An optional bold "Page N" heading can be inserted at the top of each page's text.
  • The result is a real Office Open XML .docx package, the same format Word itself saves, built with JSZip.
  • A scanned page or other image only PDF has no text layer, so the tool can't read it; a warning points you to the OCR tool instead.
  • Converts one PDF at a time, up to 100 MB.

Specifications

Specification Details
Accepts PDF files with a text layer, one at a time
Output A .docx file, A4 page size with 2 cm margins
Layout options Merge wrapped lines into paragraphs, or keep every line separate
Page headings Optional bold "Page N" heading per page
Not extracted Images, tables, columns, fonts and colours; only text is read
File size limit 100 MB
Libraries PDF.js reads the file; JSZip packages the .docx; both load from the cdnjs CDN
Where your file goes Nowhere. It stays in this tab's memory.
Good to know

What the converter does not do

It rebuilds a PDF's text as Word paragraphs. Some things a PDF does not really store cannot come back.

No scanned PDFs

A scanned page is a picture with no text layer, so the tool can't read it. Run it through PDF and image OCR first, then paste the recognised text into a document.

No images or tables

Only text comes out. Pictures and real Word tables are not extracted; a table's cells usually reappear as separate lines of text instead.

No original formatting

Fonts, colours, bold and italic styling, and multi-column layouts are not kept, and there's no real Word page break between source pages, only an optional "Page N" heading.

Reflow is a guess

Paragraph breaks come from measuring gaps between lines of text, not from the PDF's real structure, so unusual layouts can reflow oddly. Switch to "keep every line as its own paragraph" if paragraphs merge wrong.

One file at a time

The queue holds a single PDF, up to 100 MB. To go the other way, use DOCX to PDF.

Related tools

DOCX to PDF Fixed-layout PDF from a Word file, without Word. PDF and image OCR Tesseract in a worker, for scans and photos of text. PDF to images Render each page as a png or jpg image. PDF viewer Read and zoom a PDF without leaving the tab.
All document tools → Turning a scanned PDF into text with OCR →

Frequently asked questions

The converter rebuilds your PDF's text as editable Word paragraphs. Reading order and paragraph breaks carry over, but columns, tables, images and the exact font styling do not, because a PDF stores positioned characters, not document structure.

A scanned PDF is a picture of a page with no text layer, so this tool can't read it. Run the file through PDF and image OCR first to recognise the text, then paste it into a document.

The tool can't open it. Remove the password in the application that created the PDF, or with another PDF tool, and try again.

No. PDF.js opens the file and JSZip assembles the .docx package, entirely inside this browser tab. No byte of your document leaves your device.

Yes. The tool writes a standard Office Open XML WordprocessingML package, the same .docx format Word itself saves, so it opens in Microsoft Word, Google Docs, LibreOffice Writer, Pages and WPS Office.

By default, wrapped lines merge back into flowing paragraphs. This setting keeps the PDF's original line breaks instead, which suits tables, code, addresses and other content that should not be reflowed.

Your file stays on your device

PDF.js reads the file and JSZip builds the .docx, both inside this tab. Your document is not uploaded or stored, and closing the tab clears it from memory. Only those libraries are downloaded, from the cdnjs CDN.