Convert PDF to Word Online
We read the real text straight out of your PDF wherever it exists — no guessing involved. For scanned pages, we run high-accuracy OCR so the rest of the document is still fully editable.
⚡ Real text extraction + OCR fallback, not OCR on everythingDrag & drop your PDF here
or click to browse from your computer
Max 100MB per upload
Document language
Only matters for scanned pages — matching the real language noticeably improves OCR accuracy.
Enhance scanned pages
Sharpens and cleans up low-quality scans before OCR runs. Recommended.
How accuracy works here: pages with a real text layer are extracted directly and come out exact. Pages that are scanned images go through OCR, which is very accurate on clean, high-resolution scans but can still misread blurry, handwritten, or low-quality pages — we'll show you which pages used which method.
Only convert specific pages instead of the whole document.
Preparing...
Your Word document is ready
Exact where it can be
Digital text is read directly from the PDF, not re-guessed from an image, so it comes out character-for-character correct.
High-accuracy OCR fallback
Scanned pages are enhanced and run through OCR in your chosen language for the best realistic result.
Completely free
No sign-up required — convert files up to 150 pages and 100MB at no cost.
How to Convert PDF to Word
- Upload your PDF — drag and drop it, or browse your files.
- Pick a document language (only affects OCR accuracy on scanned pages) and leave "Enhance scanned pages" on for best results.
- Click Convert. Pages with real text are extracted directly; pages that are scanned images go through OCR automatically.
- Download your .docx file, ready to open and edit in Word, Google Docs, or LibreOffice.
What You Actually Get
The output is always a modern .docx file — not the legacy .doc format. Here's exactly what carries over from your PDF, and what doesn't:
Preserved
- Text — extracted character-for-character wherever a real text layer exists.
- Bold, italics, font size, and font color — carried over from the original PDF.
- Alignment and indentation — reconstructed from each line's position on the page.
- Headings — short lines with a noticeably larger font size are automatically detected and tagged as Word headings.
- Images — embedded in the document, for pages with a real text layer.
- Lists — bulleted and numbered lists become real, editable Word list formatting, not just plain-text bullet or number characters.
- Underline — preserved when a whole line is underlined; a single underlined word within an otherwise plain sentence isn't detected.
- Tables — bordered, ruled tables are rebuilt as real, editable Word table cells; tables without visible grid lines still convert as plain text.
- Font family — mapped to the closest common Word font (Arial, Times New Roman, Calibri, and similar) when recognized, though not an exact match to your PDF's original typeface.
Not Preserved
- Exact multi-column layout — columns (up to 3) are reflowed in a best-effort reading order, not pixel-perfect.
- Highlighting — highlighted or marked-up text isn't carried over.
This is genuine layout-aware reconstruction, not a plain text dump — but it's built from position and size data, not a structural copy of your PDF. If your document is table-heavy or has a highly complex layout, expect to do some manual cleanup after converting.
Real Text vs. Scanned Pages (OCR)
Every page is handled automatically based on what it actually contains — you don't need to tell the tool which pages are scanned.
- Pages with a real text layer are extracted directly. This is exact, not a guess.
- Pages with no text layer (scanned images) are run through OCR in your chosen language. OCR is very accurate on clean, high-resolution scans, but can misread blurry, handwritten, or low-quality pages.
OCR currently supports 11 languages: English, Spanish, French, German, Portuguese, Italian, Hindi, Chinese (Simplified), Arabic, Russian, and Japanese.
File Size, Page Limits, and Encrypted PDFs
| Limit | Details |
|---|---|
| File size | Up to 100MB |
| Page count | Up to 150 pages per document, for every conversion |
| Password-protected PDFs | Not supported — remove the password first |
Have a document over 150 pages? Split it into smaller files with Split PDF first, then convert each part separately.
Privacy & Processing
Conversion happens on our server — there's no client-side option for this tool, since text and layout reconstruction require server-side processing. Your file is handled entirely in memory: it exists only for the moment it takes to convert it, CurePDF's own application code never writes it to disk, and it is never used to train any AI model. No account or sign-up is required.
Frequently Asked Questions
Is CurePDF's PDF to Word converter free?
Yes — no sign-up required, and there's no charge for converting documents up to the size and page limits above.
Does it output a DOC or DOCX file?
Always a modern .docx file, compatible with Microsoft Word, Google Docs, and LibreOffice.
Will my formatting be preserved?
Bold, italics, font size, font color, alignment, indentation, headings, lists, and fully underlined lines are all preserved. Font typeface is mapped to the closest common Word font when recognized, though not an exact match to your PDF's original font.
Are tables converted into real Word tables?
Bordered, ruled tables are rebuilt as real, editable Word table cells. Tables without visible grid lines still convert as plain text with approximate spacing.
Can I convert a scanned PDF?
Yes. Any page with no real text layer automatically goes through OCR — you don't need to turn anything on manually.
How accurate is the OCR?
Very accurate on clean, high-resolution scans. It can still misread blurry, handwritten, or low-quality pages.
What languages does OCR support?
English, Spanish, French, German, Portuguese, Italian, Hindi, Chinese (Simplified), Arabic, Russian, and Japanese.
Is there a page limit?
Yes — up to 150 pages per document. This applies to every conversion, not just ones with scanned pages.
Is there a file size limit?
Up to 100MB per file.
Can I convert a password-protected PDF?
Not yet — encrypted PDFs aren't supported.
Do I need to create an account?
No sign-up is required.
Are my files stored on your servers?
No — conversion happens in memory, and CurePDF doesn't write your file to disk at the application level.
Will images in my PDF carry over?
Yes, for pages with a real text layer. Scanned pages that go through OCR become editable text only, without the original page image.
Does multi-column text convert correctly?
CurePDF makes a best-effort attempt to detect columns and reflow them in reading order, but this isn't guaranteed to be exact on complex layouts.
What happens to headings?
Short lines with a noticeably larger font size than the surrounding text are automatically detected and tagged as Word headings.
What if my PDF has more than 150 pages?
Split it into smaller files first with Split PDF, then convert each part separately.