Skip to main content

PDF to Word

PDF to Word

Transform read-only PDFs into fully editable Word documents.

Convert PDF to Word in your browser

Drop PDF files here. They are converted on your device and never uploaded.

Max 100MB per fileUp to 50 files

How to Convert PDF to Editable Word (DOCX)

How to Use PDF to Word

  1. 1

    Upload PDF Document

    Drop the PDF you want to convert into Word.

  2. 2

    Configure Layout Options

    Choose text recognition for scanned pages, and whether tables, pictures, running headers and footers, and Word’s heading and list styles are carried over.

  3. 3

    Convert & Download DOCX

    Click Convert to Word to see the document, correct anything in the editor, then download the .docx. A text PDF takes seconds; scanned pages take longer because each one is recognised.

What PDF to Word Can Do

  • Native OpenXML DOCX

    Generates standard Microsoft Word documents compatible with Office 365 and Google Docs.

  • Paragraph & Style Extraction

    Carries bold, italics, font sizes, colours, alignment, indents and heading levels into real Word styles.

  • Table Structure Retention

    Converts PDF tables into native editable Word tables, with column widths measured from the page.

  • Zero File Uploads

    Convert confidential contracts locally without cloud conversion servers.

Tools & Settings Reference

Preserve Layout
Keeps headings, bullets and numbered sequences as real Word styles. Switch it off and everything becomes plain paragraphs, with list markers written into the text so nothing is lost.
Text Recognition (Auto / Text Layer Only / Force OCR)
Auto reads the text layer and only recognises pages that have none, in the language you pick (55 offered). Text Layer Only never runs OCR, so a scanned page has no text in the Word file; its picture is kept when Extract images is on. Force OCR re-reads every page from its pixels, which is what you want when the existing text layer is itself garbled.
Detect Tables
Reconstructs aligned columns into real Word tables instead of tab-separated lines.
Running Headers & Footers
Text repeated at the top or bottom of most pages, page numbers included, goes into Word’s own header and footer, with a page number Word keeps up to date, instead of repeating through the body.
Target Font & Line Spacing
Applied when the document is written, so changing them only needs a fresh download, not a fresh parse.

Frequently Asked Questions

The table becomes a real, editable Word table: PDFixel reads cell bounding boxes and writes native OpenXML structures (`w:tbl`, `w:tr`, `w:tc`), with column widths measured from the page, and cells merged across columns are rebuilt as merged cells. Cells merged vertically down a column are rebuilt as separate cells, so those need a manual pass afterwards.

Yes, directly. Leave Text Recognition on Auto and any page without a text layer is recognised before it is written, so the .docx contains editable text rather than a blank page. The Report tab then shows the recognition confidence, which is the number to check before trusting the result — a low score means a low-resolution or skewed scan and you should expect wrong characters.

With Preserve Layout on, yes: lines are grouped by spacing and font size, then written as Word paragraphs with real Heading 1/2/3 styles and genuine bulleted and numbered lists. Headings are inferred from type size and weight, not read from the PDF, so an unusually styled document can need a correction or two. With Preserve Layout off everything is written as plain paragraphs and bullet characters are kept inline instead.

A PDF that only restricts printing or copying converts as it is. One that needs a password to open is refused here: unlock it first with Unlock PDF, which needs that password, then convert the unlocked copy.

The PDF has a text layer that is itself wrong — usually a bad earlier OCR pass, or fonts whose character mapping is broken, which is common in files produced by cheap scanners and some report generators. Extraction faithfully copies out what is there, so the output reads as gibberish. Set Text Recognition to Force OCR: it ignores the text layer entirely and re-reads every page from the pixels.

Related Guides

Related Tools