Convert PDF to DOCX
PDF text to editable Word, in your browser.
Drop your .pdf file
or click to browse · max 50MB
or paste with ⌘V
Converting a PDF to DOCX extracts the text from your PDF and packages it into an editable Word document. It's the right move when you need to copy, edit, or repurpose the contents of a PDF without retyping. Our converter detects headings, multi-column rows, and tables — so an invoice's table grid or a report's heading hierarchy comes through, not just flat text. Embedded images and exact font matching aren't preserved; for pixel-perfect fidelity you'll want a desktop tool like Adobe Acrobat or Microsoft Word's PDF import.
When you'd use it
Edit a contract or letter
Got a PDF you need to make small edits to? Converting to DOCX gives you an editable starting point. Faster than retyping, far more reliable than copy-pasting from a PDF reader.
Repurpose article content
Pulling text out of a PDF report or paper to use elsewhere — a blog post, a newsletter, an internal doc — is much faster as DOCX than as raw clipboard text.
Strip styling for a clean baseline
Sometimes the goal isn't to preserve the PDF's look but to get the words into Word so you can re-style them with your own template. This conversion delivers that.
How it works
-
Your file never leaves the browser.
-
Conversion runs locally via WebAssembly and Canvas.
-
Download. No account, no history.
Technical notes
This converter uses pdf.js to extract text-content tuples (string, x, y, font, size) from each page, then runs a four-stage layout reconstruction: (1) group text items into rows by Y-coordinate, (2) cluster items within each row into columns by X-gap, (3) detect heading rows by font-size ratio against the document median, (4) detect multi-row tables when consecutive rows share the same column structure, emit them as real Word tables. Bold and italic runs are inferred from the embedded font name. Honest limits remaining: embedded images aren't extracted, exact font/color matching isn't preserved (the docx is monochrome), and academically-formatted multi-column flowing prose collapses into row-paired tables rather than recombining into single columns — that last one is a hard reading-order problem that no in-browser tool currently solves.
Frequently asked
What happens to my PDF files?
Nothing leaves your browser. The conversion runs locally on your device using WebAssembly and the Canvas API. No file is uploaded to a server — including ours.
Is the converted DOCX identical to the original?
DOCX is a lossless format. Pixel data round-trips exactly when the input is also lossless.
Is there a file size limit?
There is no hard cap, but very large files (hundreds of MB) may be slow on older devices because everything happens in your browser's memory.
Do you collect data on my file or my use?
Your file is invisible to us — it never leaves your browser. For aggregate page-view counts we use Cloudflare's server-side edge logs (no JavaScript beacon runs in your browser). We display Google AdSense ads on conversion result pages; the AdSense script loads on every page and may set advertising cookies after you accept the consent banner (Google Consent Mode v2 asks first). The ads see the page URL and standard browsing signals — they never see your file, because the file never leaves your device.
Will the layout look exactly like the PDF?
Close, not exact. Headings come out as Word heading styles, tables as Word tables (editable cell-by-cell), and multi-column header rows stay aligned. What's lost: exact font choices, colors, embedded images, and the precise positioning of free-floating elements. If you need pixel-perfect fidelity (logos, brand colors, exact spacing), use Adobe Acrobat or Microsoft Word's built-in PDF import (Word 2016+).
What about scanned PDFs?
Scanned PDFs (where each page is just an image of text) won't extract text because there's no text layer to extract. You'd need OCR first. For now this tool only handles text-layered PDFs — the kind exported from Word, web pages, or modern document scanners with built-in OCR.
Why is some text in the wrong order?
PDFs store text in render order, not reading order. Our converter groups text by vertical position and clusters columns by horizontal gaps, which works well for invoices, reports, and forms — but academic-style two-column flowing prose (where a paragraph wraps across both columns of one page) is genuinely hard. We pair the columns row-by-row so they stay readable side-by-side, but you may need to manually splice the columns together for true reading order. This is a research problem; no in-browser tool currently solves it.
Are there file-size limits?
No hard cap, but PDFs over ~50 pages can be slow to process in the browser. If your file is larger, consider splitting it (we don't have that tool yet) or using a desktop application.