How do I convert a PDF to an editable Word document?

Choose a PDF and conversion starts immediately — there is no button to press. Every page is read, paragraphs are rebuilt, and the finished .docx downloads on its own when it is ready. A progress line shows which page is being processed, and a Re-convert button lets you run it again.

What formatting survives the conversion?

More than most in-browser converters manage: bold and italic runs, three levels of heading detected from font size, centred and right-aligned lines, bulleted and numbered lists including nested levels, page breaks between pages, and words rejoined across hyphenated line breaks. Font sizes are kept where they differ from the body text.

Are tables preserved?

No — this is the main limitation. There is no table reconstruction, so cell contents come through as ordinary lines of text and the grid itself is lost. If the table is the point of the document, convert to a spreadsheet with the PDF to Excel tool instead, which is built for tabular structure.

Does it handle two-column layouts without jumbling the text?

Yes. Naive extraction reads straight across the page and interleaves both columns line by line, producing nonsense. This tool detects a consistent empty gutter with substantial text either side, then reads each column top to bottom in turn. Single-column pages, the vast majority, are untouched by that logic.

What happens with a scanned PDF that has no real text?

Each text-free page is rendered to a sharp 2× image and embedded in the Word file as a picture. You get a valid .docx, but those pages are photographs of text, not editable words. Run the PDF through the OCR tool first if you need to edit the content.

Can I convert a password-protected PDF?

No. An encrypted PDF cannot be opened for reading, and the tool reports the failure rather than producing a broken file. Remove the password in a PDF reader that has the credentials, save an unprotected copy, then convert that.

Is there a file-size or page limit?

No artificial cap — the real limit is your device memory, since the whole document is processed in the tab. Text-heavy PDFs are fast. Scanned files are much heavier because every page is rendered to a 2× bitmap, so a long scan on a phone is the case most likely to struggle.

Is my PDF uploaded anywhere, and does the output carry a watermark?

Nothing is uploaded. Text extraction and .docx assembly run entirely in your browser, so confidential contracts, payslips and reports never leave the device. Output is a standard OpenXML file with no watermark, no page cap and no account required. Building .docx needs the Compression Streams API, so use a current browser.

PDF to Word

Convert PDF to Word (DOCX) online free — editable document with preserved formatting, fonts and tables. In-browser processing — no upload, no account.

Files never leave your deviceInstant — no upload waitFree, no sign-up

Choose a PDF to convert to Word or drag & drop100% private · all pages · bold, italic, headings, lists & alignment preserved

or import from

PDF to Word rebuilds a PDF as an editable .docx file. This one runs entirely in your browser — no upload, no sign-up, no watermark — and reconstructs headings, bold and italic text, lists, alignment and two-column layouts rather than dumping a wall of plain text. Conversion starts the moment you choose a file.

How the conversion works

Choose a PDF from your device or Google Drive. Processing begins straight away.
Watch the page counter — each page is read, its text items sorted into lines, and lines grouped into paragraphs.
The .docx downloads automatically when the last page is done. Re-convert if you want to run it again.

The output is standard OpenXML, so it opens in Microsoft Word, Google Docs, LibreOffice and Apple Pages without conversion warnings. Note that building a .docx requires the browser's Compression Streams API — current Chrome, Edge, Firefox and Safari all have it, but a very old browser will be told plainly rather than failing silently.

What actually survives the conversion

Most in-browser converters give you unstyled paragraphs. This one infers structure from the geometry of the text on each page, which recovers considerably more:

Formatting carried into the Word document, and what is not.
Element	Result	How it is worked out
Bold and italic	Preserved	Read from the embedded font name of each run
Headings	Three levels	Font height against the page's most common body size
Centred / right-aligned lines	Preserved	Line midpoint compared with the page centre
Bullet and numbered lists	Real Word lists, nested	Leading marker detected, indentation clustered into levels
Paragraph and section breaks	Preserved	Vertical gaps measured against the median line spacing
Page breaks	Preserved	One per source page
Hyphenated line breaks	Words rejoined	Trailing hyphen removed and the line merged
Two-column layouts	Read column by column	Consistent empty gutter detected across the page
Tables	Grid lost	No table reconstruction — cells become text lines
Inline images on text pages	Dropped	Only text runs are extracted

Digital PDFs, scanned PDFs, and what you get from each

This distinction decides whether the result is editable at all. A digital PDF — exported from Word, Docs, LaTeX or a print-to-PDF driver — contains real text objects, and those convert cleanly. A scanned PDF from a flatbed or phone camera contains only pixels; there is no text to extract.

Rather than skipping those pages, the tool renders each text-free page at 2× and embeds it in the Word file as a picture, sized to the A4 content width. You get a complete document with every page present and in order, which is genuinely useful for archiving or for mixed files where only a few pages are scans — but the scanned pages are images of words, not words. To edit them, run the file through OCR PDF first to add a real text layer, then convert that.

Where the heuristics can trip up

Structure is inferred, not read from a stylesheet, so a few honest failure modes are worth knowing.Numbered lists are always rendered as decimal — a source list using "(a)", "(b)" or roman numerals becomes 1., 2., 3. in Word. A line that begins with a short word and a full stop can be mistaken for a list marker, so a paragraph opening "Mr. Patel agreed…" may arrive as a list item. Heading detection depends on the document actually using larger type for headings; a report that signals headings with bold-only text at body size will keep the bold but not the heading level.

None of these corrupt your content — they cost a minute of cleanup in Word. The trade is a conversion that never uploads the document, which for a contract or a payslip is usually the trade you want.

How it compares to cloud converters and Acrobat

Converting one confidential PDF, three ways.
	ToolJiffy	Typical online converter	Desktop Acrobat
Document uploaded to a server	No	Yes, and queued on disk	No
Sign-up	No	Often after one free file	Subscription
Page or size limit	Device memory only	Commonly capped free	None practical
Watermark on output	Never	Sometimes on free tiers	No
Table reconstruction	No	Sometimes	Yes
Scanned pages	Embedded as images	Varies; OCR often paid	Built-in OCR
Works offline	Yes, once loaded	No	Yes
Price	Free, unlimited	Free tier then paid	Paid

Related PDF tools

If the document is mostly tables, PDF to Excel is the right target andPDF to PowerPoint handles decks. For scans,OCR PDF adds a searchable text layer in place. To take pages out before converting, use extract pages orsplit PDF — converting three pages instead of ninety is faster and leaves less to tidy up. Going the other direction, Word to PDF closes the loop.

Frequently Asked Questions

How do I convert a PDF to an editable Word document?
Choose a PDF and conversion starts immediately — there is no button to press. Every page is read, paragraphs are rebuilt, and the finished .docx downloads on its own when it is ready. A progress line shows which page is being processed, and a Re-convert button lets you run it again.
What formatting survives the conversion?
More than most in-browser converters manage: bold and italic runs, three levels of heading detected from font size, centred and right-aligned lines, bulleted and numbered lists including nested levels, page breaks between pages, and words rejoined across hyphenated line breaks. Font sizes are kept where they differ from the body text.
Are tables preserved?
No — this is the main limitation. There is no table reconstruction, so cell contents come through as ordinary lines of text and the grid itself is lost. If the table is the point of the document, convert to a spreadsheet with the PDF to Excel tool instead, which is built for tabular structure.
Does it handle two-column layouts without jumbling the text?
Yes. Naive extraction reads straight across the page and interleaves both columns line by line, producing nonsense. This tool detects a consistent empty gutter with substantial text either side, then reads each column top to bottom in turn. Single-column pages, the vast majority, are untouched by that logic.
What happens with a scanned PDF that has no real text?
Each text-free page is rendered to a sharp 2× image and embedded in the Word file as a picture. You get a valid .docx, but those pages are photographs of text, not editable words. Run the PDF through the OCR tool first if you need to edit the content.
Can I convert a password-protected PDF?
No. An encrypted PDF cannot be opened for reading, and the tool reports the failure rather than producing a broken file. Remove the password in a PDF reader that has the credentials, save an unprotected copy, then convert that.
Is there a file-size or page limit?
No artificial cap — the real limit is your device memory, since the whole document is processed in the tab. Text-heavy PDFs are fast. Scanned files are much heavier because every page is rendered to a 2× bitmap, so a long scan on a phone is the case most likely to struggle.
Is my PDF uploaded anywhere, and does the output carry a watermark?
Nothing is uploaded. Text extraction and .docx assembly run entirely in your browser, so confidential contracts, payslips and reports never leave the device. Output is a standard OpenXML file with no watermark, no page cap and no account required. Building .docx needs the Compression Streams API, so use a current browser.