Skip to content

XPS / OpenXPS to PDF converter

XPS → PDF rescues a file that someone printed to Microsoft XPS Document Writer: the pages are redrawn as PDF with the fonts the XPS already carries, so nothing is retyped or reflowed.

  • Converted with Image engine · input deleted within 24 h
  • Free: 5 conversions a day · 5 MB each
  • Higher limits →

From real runs

What you get — and what changes

  • Page size and page count Kept

    A three-page Document Writer print-out stayed three A4 pages at 594.75 × 841.5 pt and a PowerPoint deck saved as XPS stayed three 720 × 540 pt slides. XPS counts in 1/96 inch, which is why its A4 is half a point narrower than the usual 595.28.

  • Fonts Kept

    An XPS package holds its own copies of the fonts it uses. Calibri, Cambria, Segoe UI, Consolas and MS Gothic all came out embedded in the PDFs, so no typeface is swapped and no line wraps differently.

  • Text Kept

    All 8,224 characters of the Go game record, Japanese player names included, could be searched and copied from the PDF, the same count the XPS reported.

  • Lines, shading and charts Kept

    Table rules, grey row bands and the Ghostscript colour-wheel test page stay vector drawings, so they remain sharp when you zoom or print.

  • Photos and icons Kept

    The JPEG photo inside a Word-made XPS was copied into the PDF unchanged, 3,200 bytes in and 3,200 out; a PNG icon kept its transparency as a separate mask.

  • Web links Kept

    The link in the Tika test document still opens its web address in the PDF. A plain XPS-to-PDF rebuild drops links, so our image engine copies them across page by page.

  • Title and author Kept

    When the package has a core-properties part, as Office's Save as XPS writes, its title and author become the PDF's document properties. Document Writer print-outs usually have none to copy.

  • Arabic and other right-to-left text Changed

    Arabic lines look right on the page, but copying them yields Arabic presentation-form characters (U+FExx), which some search boxes and translation tools do not match against normal Arabic.

  • Print settings and thumbnail Lost

    The PrintTicket files Windows adds (tray, duplex, copies) and the package thumbnail are ignored; the PDF holds the pages only.

Sample conversions

Three-page print-out from XPS Document Writer

Input
76,412-byte .xps of a Go game analysis printed from a Windows program: A4, a 60-row table on each page, two embedded Japanese fonts.
Output
56,999-byte PDF, 3 pages, in 36 ms. Rules, row shading and names match the XPS; MS Gothic and MS UI Gothic are embedded.

OpenXPS file from Windows 10

Input
51,175-byte OpenXPS document, the flavour Windows 8 and later write by default, with one A4 page of Consolas text and streamed ZIP headers.
Output
45,695-byte PDF in 18 ms with selectable text and Consolas embedded. The same file renamed from .xps to .oxps gave an identical result.

Word document saved as XPS

Input
560,946-byte XPS from Word with Arabic and English lines, a web link, a photo, an attachment icon and five typefaces.
Output
566,317-byte Letter-size PDF in 130 ms. The link works, the photo is untouched and the author from the package fills the PDF author field; copied Arabic comes out as presentation forms.

A real XPS / OpenXPS → PDF run

We converted our XPS / OpenXPS sample (74 KB) with Image engine on 2026-10-20. It took 21 ms and produced a 67 KB PDF file, about the same size as the input. The engine raised no warnings.

Input · sample.xps · 74 KB

ZIP with 18 parts

Archive parts
18
first
[Content_Types].xml

Output · sample.pdf · 67 KB

PDF 1.7, 3 pages, selectable text (Calibri), 0 image objects

Version
1.7
Pages
3
Fonts
Calibri
Images
0
Text layer
yes

Options and when to change them

max_pages
Set it when you only need the first pages of a long print-out, for example 1 for a receipt or cover sheet.

Do it yourself

MuPDF (mutool)

bash
mutool convert -O compress -o out.pdf document.oxps

Reads both .xps and .oxps. With mutool 1.21 the web link in our test file did not survive, and without -O compress the PDF was three times larger than the XPS.

Windows

bash
Open the file in XPS Viewer → Print → Microsoft Print to PDF

XPS Viewer is an optional Windows feature that recent installs leave out; printing also flattens links.

XPS / OpenXPS to PDF by API or MCP

The HTTP API and the MCP server run the exact engine this page uses (Image engine); the response carries the same warnings the web converter shows. Check support first with GET /api/v1/pairs?from=xps&to=pdf.

curl
curl -X POST https://filetypeconverters.com/api/v1/convert \
  -H "Authorization: Bearer $FTC_API_KEY" \
  -F "file=@input.xps" -F "from=xps" -F "to=pdf" \
  -F 'options={"width":0}' \
  --output output.pdf

Options this engine accepts for XPS / OpenXPS → PDF

KeyApplies toTypeDefault
widthWriting PDFnumber0
heightWriting PDFnumber0
fitWriting PDFcontain · cover · fillcontain
qualityWriting PDFnumber90
backgroundWriting PDFstring#ffffff
stripMetadataWriting PDFbooleantrue
API reference

When this is the wrong conversion

  • If you still have the program that printed the XPS, print again and choose Microsoft Print to PDF; that skips one conversion.
  • If you need to change the wording, a PDF is not editable either; convert to PDF first, then use a PDF to Word conversion, and expect tables to need tidying.
  • If the file only needs to go into a ticket or chat as a picture, XPS → PNG gives one image per page.

XPS / OpenXPS to PDF questions

What is the difference between .xps and .oxps?

OXPS is OpenXPS, the ECMA-388 standard version that Windows 8 and later save by default; .xps is the original Microsoft format from Vista and Windows 7. They use different XML namespaces but describe pages the same way, and both convert here with the same result.

Why can't my Mac or phone open an .xps file?

macOS, iOS and Android have no XPS viewer, and Windows stopped installing one by default around 2018. PDF opens on all of them, which is why converting is usually the quickest way to read an XPS someone sent you.

The converter says my file is a Word or EPUB file, not XPS. Why?

Every real XPS package contains a FixedDocumentSequence part that points to its pages. A .docx, .pptx or .epub renamed to .xps has none, so we stop before converting and say what the file actually is; open it in the matching program instead.

Will the PDF look different from the XPS on screen?

It should not. The XPS already contains its fonts and exact glyph positions, so the PDF reuses them; in our tests every page kept its size, line breaks and table layout.