PDF to Word

What is PDF to Word?

PDF and Word represent the same content with completely different mental models. PDF places every glyph at an explicit (x, y) coordinate. Word flows characters through paragraph styles and lets the renderer decide line breaks. The conversion is non-trivial: a "good" PDF→Word converter has to reverse-engineer flowing paragraphs from absolute glyph positions, recognise that visually-adjacent text in two columns isn't one continuous sentence, infer table cell boundaries from whitespace patterns, and decide whether a horizontal line is a table border or a decorative rule. This tool runs the conversion server-side via LibreOffice headless — the same engine that powers Collabora Online and most enterprise PDF→DOCX backends. LibreOffice is the only freely-available converter that handles all of: multi-column academic paper layouts, complex tables with merged cells, embedded fonts (preserved verbatim in the output DOCX), and footnotes with reference markers. The result is a real .docx — not the "RTF renamed to .docx" that some free converters output and then mysteriously breaks when re-saved.

Why use this tool?

Output is a real .docx — XML-based, MS Office–compatible, opens in Word 2007+ / Google Docs (with a file upload) / LibreOffice Writer / Apple Pages. Fonts referenced in the PDF are embedded into the .docx when the licence allows; substituted otherwise (and the substitution is logged). Tables become Word tables with editable cells. Headings detected from text size and position become Heading 1/2/3 styles you can use in Word's navigation pane. For scanned PDFs (image-only, no text layer), the conversion automatically routes through OCRmyPDF first — Tesseract 5 with the eng+osd model gives ~97% character accuracy on clean 300 DPI scans, less on phone-camera shots with lighting issues. The OCR pass adds 10–30 seconds depending on page count. Upload goes over HTTPS to our private converter, the file is deleted from disk the moment the response is sent, and we don't log filenames. No daily cap.

Common use cases

The recurring pattern: a vendor sends you a contract or supplier agreement as PDF, you need to redline 3 clauses and send it back. Or you receive a research paper and want to quote 4 paragraphs into your own document. Or HR sends a benefits enrollment PDF and wants it filled in digitally. In each case the PDF was probably exported from a Word document originally — converting back recovers the editable form. It's also the right tool when you have a PDF you never had the .doc/.docx source for: an old report whose original file was lost; a presentation handout where the deck is gone; a regulatory document you need to summarise for your own brief. For these, expect to spend 2–5 minutes cleaning up minor layout differences after conversion — that's normal for any PDF→Word path, not a tool defect.

How to use PDF to Word

1. 1Drop your PDF onto the upload area — text and scanned PDFs both work
2. 2The converter detects whether the PDF has a text layer; if not, OCR runs automatically (adds 10–30 s for scans)
3. 3LibreOffice headless handles the layout reconstruction — typically 3–8 s for text PDFs, 15–60 s for scans
4. 4Download the .docx — the file name is your original with the .pdf swapped for .docx
5. 5Open in Word / Google Docs / LibreOffice and edit normally. Expect 2–5 minutes of cleanup for complex layouts.

Frequently asked questions

My academic paper came out with the column flow broken — paragraphs jumping from column 2 back to column 1 mid-sentence. Why?

That's the hardest case for any PDF→Word converter. The PDF has no concept of "column 1 then column 2"; it just has glyphs at coordinates. LibreOffice usually gets column flow right when the layout uses standard multi-column section formatting; it fails when the original was set with floating text boxes or absolute positioning (common in InDesign exports). The workaround is to convert page-by-page (split the PDF first with /pdf/split-pdf into 1-page chunks, convert each, and assemble in Word), or accept the manual cleanup.

The output .docx is huge — 80 MB for a 30-page PDF that was 6 MB. What happened?

The converter embeds the source PDF's fonts and images into the DOCX. If the PDF had high-resolution embedded images (a scanned page is essentially a 300 DPI JPEG per page), those land in the .docx at the same resolution. Run the result through Word's File → Reduce File Size, or replace the embedded images with downsampled versions if you don't need print-quality.

Does it work for forms — can I keep the form fields editable?

Partially. PDF AcroForm fields (text inputs, checkboxes) are converted to corresponding Word content controls when LibreOffice recognises them, but the round-trip isn't lossless — calculated fields, JavaScript actions, and dynamic XFA forms don't survive. For forms specifically, /pdf/fill-pdf is usually a better fit than converting to Word.

Can I convert a password-protected PDF?

No — LibreOffice refuses encrypted PDFs by design. Unlock first with /pdf/unlock-pdf (you need the password), convert, then re-protect the output .docx in Word if needed (Word natively supports document password encryption).

How does this compare to Adobe Acrobat's PDF→Word conversion?

Acrobat is broadly more accurate on complex layouts because Adobe owns the PDF spec and has been refining their converter for 20 years. For typical business documents (contracts, reports, articles, resumes), the difference is small to invisible. For research papers with footnotes, equations, and multi-column flow, Acrobat wins materially — if that's your use case, the $15/month subscription pays back in cleanup time. For everything else, LibreOffice is good enough and free.

My scanned PDF came back with text that's mostly right but with weird characters in places. Why?

Tesseract OCR is reading bitmap pixels and guessing characters. Confusion between "O" and "0", "l" and "1", "rn" and "m" is common — especially with anti-aliased low-resolution scans. For best OCR results: scan at 300 DPI minimum, ensure the page is straight (no skew), and use a flatbed scanner not a phone camera if accuracy matters. Run /image/sharpen-image on each page before re-converting if the scans are soft.

Does the .docx preserve hyperlinks from the PDF?

Yes — explicit hyperlinks (text or button annotations linked to URLs) carry through as clickable links in the Word output. Implicit URLs (plain text like "https://example.com" without a link annotation) become clickable in Word automatically when Word detects them on save.

1. 1If your PDF has a known structure (a contract template, a recurring report), do the conversion once, then save the cleaned-up .docx as your master template. Future versions of the same PDF will convert with the same layout quirks, so your cleanup work pays back across all of them.
2. 2For PDFs with mostly text but a few problematic tables, convert here and then ALSO run /pdf/pdf-to-excel for the table pages — Excel's table-extraction path is more accurate than Word's. Paste the cleaned Excel table back into the Word doc.
3. 3After conversion, run Word's Spelling & Grammar pass even if the source PDF was clean. OCR'd scans frequently produce "real" misspellings (substituting valid words like "ot" for "of") that Word will catch but a visual scan won't.

How does it compare?

Adobe Acrobat Pro's PDF→Word is the accuracy gold standard at $14.99/month minimum. Nitro PDF is comparable at similar pricing. Free competitors: Smallpdf and iLovePDF cap free use to 2 conversions per day; Google Drive's PDF→Docs conversion runs in the cloud and frequently mangles tables. This tool gives you Acrobat-class engine quality (LibreOffice is what most enterprise PDF servers actually run under the hood) with no daily cap, no watermark, and a documented retention policy.

Related guides

Editorial walk-throughs that go deeper on the workflow most people use this tool for.

Related tools

More free tools. No signup, no watermarks.