Skip to main content

DOCX to PDF converter

Convert Word documents to PDF

No watermarkFree & unlimited
All processing happens in your browser. No data is sent to any server.

A .docx is a ZIP of XML parts, and this page opens it with Mammoth, a JavaScript library that maps Word paragraph and character styles onto semantic HTML. That HTML is laid out in a hidden element sized to a 210 mm column, html2canvas photographs it at double resolution, and jsPDF writes the capture into A4 pages with 15 mm margins. The result is a faithful picture of the document flow rather than a re-typeset copy: the text is rendered, not embedded, so treat the output as a read-only rendition rather than an editable or searchable PDF.

Key facts about DOCX to PDF converter

Key facts about DOCX to PDF converter
ReaderMammoth 1.12 maps Word styles onto HTML: headings, lists, tables, bold and italic, inline images, footnotes.
Text in the PDFRasterised, not embedded. The pages contain a JPEG at quality 0.98, so you cannot select, search or copy text.
FontsNot carried over. The render is forced to Arial at 12 pt with 1.6 line spacing, whatever the document specified.
Page geometryA4 portrait, 15 mm margins on all sides, regardless of the section setup inside the Word file.
Headers and footersDropped. Mammoth does not emit them, so running heads and footer text never reach the render.
Page numbersWord field codes for page numbers produce nothing, because the HTML stage has no concept of pages.
Page breaksThe capture is one tall bitmap sliced to A4, so a break can land mid-line or mid-row of a table.
HyperlinksVisible as styled text but not clickable: rasterising the page removes the link annotations.
ImagesEmbedded pictures are inlined by Mammoth as data URIs and appear in the capture at their HTML size.
Protected filesA password-protected .docx cannot be opened. Remove the protection in Word first.
Capture resolutionhtml2canvas runs at scale 2, roughly 192 DPI against the 210 mm column, which is screen quality rather than press.
File ceilingThe picker rejects anything above 500 MB, and up to twenty documents can be queued in one run.

What happens to your file

The document is read with file.arrayBuffer() and unzipped by Mammoth inside this tab, so its contents are parsed on your own machine. The intermediate HTML is attached to the page just long enough for html2canvas to photograph it, then removed from the DOM. jsPDF builds the PDF in memory and hands back a Blob that becomes a local download link. No part of the file, and no extracted text, is sent to any server: there is no upload endpoint behind this tool and therefore no stored copy to worry about.

About this tool

  1. 1

    Choose the .docx

    Drag the Word file onto the panel or click to browse. Only the modern .docx package is accepted, not legacy .doc.

  2. 2

    Clean the document first

    Accept or reject tracked changes and resolve comments in Word, so the render reflects the final text.

  3. 3

    Convert

    Mammoth extracts the content, the page renders it and jsPDF writes A4 pages. Long documents take a moment.

  4. 4

    Check the page breaks

    Scroll the result before sending it on. Breaks fall where the bitmap is sliced, not where Word put them.

  5. 5

    Download

    Save the PDF, or take a batch of converted documents as a single ZIP assembled in the page.

Specs & compatibility
Input and outputTakes .docx (the Office wordprocessingml type) and returns the PDF as a blob with the application/pdf type, named after the file you dropped in.
EngineMammoth 1.12.0 converts the .docx to HTML, and html2pdf.js 0.14.0 rasterises that HTML with html2canvas at scale 2 into an A4 jsPDF document.
ModesFile mode only: DOCX to PDF has no paste box, so the DOCX arrives through the picker or a drag and drop and the PDF comes back as a download.
Batch and capsTwenty DOCX files per run at 500 MB each; several PDF results are zipped in page memory as converted-pdf-files.zip.
Browser and deviceThe PDF comes from rasterising a hidden DOM node, so the ceiling on DOCX to PDF is device memory, not a browser version - every current browser runs it and phones run out first on a long document.
  • Use Word heading styles rather than manually enlarged bold text. Mammoth works from styles, so a paragraph styled Heading 2 converts correctly while a hand-formatted one does not.
  • Wide tables are the usual failure. The render column is 210 mm, so anything designed for landscape gets squeezed; split the table or reduce its columns before converting.
  • If you need selectable text in the PDF, export from Word itself. This route always produces page images, which is fine for distribution but not for search or accessibility.
  • Headers, footers and page numbers do not survive. Move anything essential, such as a document reference or a confidentiality note, into the body before converting.
  • Very long documents build one tall capture in memory before any page is written, so convert a 200-page manuscript on a desktop rather than a phone.
  • Produce a fixed rendition of a proposal or contract so the recipient cannot casually reflow or edit the wording.
  • Turn a Word draft into a PDF on a machine where Office is not installed and company policy forbids uploading the file anywhere.
  • Batch a folder of Word templates into PDFs for distribution, without a single document leaving the laptop.
No. The conversion renders the document to a bitmap with html2canvas and places that image inside the PDF pages, so there is no text layer to select, search, copy or read aloud. This is the trade-off for converting entirely in the browser without a print pipeline. When selectable text matters, export to PDF from Word, LibreOffice or Google Docs instead.
They are not. The intermediate HTML is styled with Arial at 12 pt and 1.6 line spacing, so a document set in Garamond or Calibri is redrawn in Arial. Mammoth deliberately extracts structure rather than direct formatting, which is why headings, lists and tables survive as structure while typeface, point size and colour do not.
Mammoth converts the document body and does not emit headers or footers, and the HTML stage has no pages for a page-number field to count. Anything that lived in the running head, including document references, dates and confidentiality markings, is absent from the render. Move that content into the body of the document before converting if it has to appear.
No. The picker accepts only .docx, the ZIP-and-XML package introduced with Office 2007, because Mammoth reads that structure. The binary .doc format is a completely different container. Open the file in Word, LibreOffice Writer or Google Docs and save it as .docx first, then convert the result here.
The document is captured as a single tall image and then sliced into A4 sheets, and the slicer has no knowledge of rows, paragraphs or widow control. A break therefore falls wherever the page boundary lands. Shortening the table, or splitting it into smaller tables separated by headings, is the practical workaround.
No. An encrypted .docx cannot be unzipped without the password, and this tool has no place to enter one, so the conversion fails at the reading stage. Open the file in Word, remove the protection under document security settings, save an unprotected copy, and convert that copy here.

Related tools

View all

Updated

We use anonymous analytics to improve ToolChamp. No personal data is stored or sold. Privacy Policy