docx-rs

Fast DOCX-to-PDF conversion with a Rust layout engine whose output matches LibreOffice's, without running LibreOffice.

from docx_rs import convert, convert_bytes

convert("document.docx", "document.pdf")
pdf_bytes = convert_bytes(open("document.docx", "rb").read())

Conversions release the GIL, so threads convert in parallel.

Supported documents

The renderer only produces a PDF when it supports everything in the document. A document using unsupported features raises docx_rs.UnsupportedDocument, whose findings list each unsupported construct; it never returns an approximate PDF. Support currently covers typical generated legal documents: styles, numbering, tables, sections, headers and footers, fields such as cached tables of contents, text boxes and images. Use docx_rs.inspect(path) to see the eligibility report without converting.

Fonts

The package bundles open fonts and substitutes them for common Microsoft fonts: TeX Gyre Termes for Times New Roman, Carlito for Calibri, Caladea for Cambria, Liberation Sans and Mono for Arial and Courier New, OpenSymbol for Symbol and DejaVu Sans for Segoe UI Symbol. These are the substitutions a typical Linux LibreOffice installation makes. Each font keeps its own licence; see docx_rs/fonts/licenses/. A custom font profile can be passed as font_profile= (a profile JSON path or docx_rs.FontProfile.load(path)).