office2pdf

Pure-Rust library and CLI for converting DOCX, XLSX, and PPTX files to PDF.

No LibreOffice, no Chromium, no Docker — just a single binary powered by Typst.

Features

• DOCX — paragraphs, inline formatting (bold/italic/underline/color), tables, images, drawing shapes, ordered/nested lists, syntax-highlighted code, headers/footers, page setup
• PPTX — slides, text boxes, shapes, tables (with theme-based table styles), images, slide masters, speaker notes, gradient backgrounds, shadow/reflection effects
• XLSX — sheets, cell formatting, merged cells, column widths, row heights, conditional formatting (DataBar, IconSet, and formula rules)
• PDF/A-2b — archival-compliant output via --pdf-a
• Embedded font extraction — fonts embedded in PPTX/DOCX are automatically extracted, deobfuscated, and used during conversion
• macOS Office font auto-discovery — PowerPoint/Word/Excel bundled fonts and Office cloud font caches are searched automatically
• WASM — runs in browsers and Node.js via WebAssembly, with optional caller-provided or feature-gated Simplified Chinese fonts
• Zero external dependencies — runs as a standalone executable

Installation

Library

[dependencies]
office2pdf = "0.6.7"

CLI

cargo install office2pdf-cli

Prebuilt binaries

Every GitHub release ships standalone CLI binaries — no Rust toolchain needed:

On Linux and macOS, download, extract, and place the binary on your PATH:

VERSION=v0.6.7
TARGET=x86_64-unknown-linux-gnu  # pick your platform's target from the table above
curl -L "https://github.com/developer0hye/office2pdf/releases/download/${VERSION}/office2pdf-${VERSION}-${TARGET}.tar.gz" | tar xz
sudo install "office2pdf-${VERSION}-${TARGET}/office2pdf" /usr/local/bin/

On Windows, unzip the archive and add office2pdf.exe to your PATH.

The macOS binaries are not notarized. Binaries downloaded with a browser are quarantined by Gatekeeper; clear the flag with xattr -d com.apple.quarantine office2pdf (downloads via curl are unaffected).

Quick Start

As a library

// Simple one-liner
let result = office2pdf::convert("report.docx").unwrap();
std::fs::write("report.pdf", &result.pdf).unwrap();

// With options
use office2pdf::config::{ConvertOptions, PaperSize};

let options = ConvertOptions {
    paper_size: Some(PaperSize::A4),
    ..Default::default()
};
let result = office2pdf::convert_with_options("slides.pptx", &options).unwrap();
std::fs::write("slides.pdf", &result.pdf).unwrap();

// In-memory conversion
use office2pdf::config::Format;

let docx_bytes = std::fs::read("report.docx").unwrap();
let result = office2pdf::convert_bytes(
    &docx_bytes,
    Format::Docx,
    &ConvertOptions::default(),
).unwrap();
std::fs::write("report.pdf", &result.pdf).unwrap();

CLI

# Single file
office2pdf report.docx

# Explicit output path
office2pdf report.docx -o output.pdf

# Batch conversion
office2pdf *.docx --outdir pdfs/

# With options
office2pdf slides.pptx --paper a4 --landscape
office2pdf spreadsheet.xlsx --sheets "Sheet1,Summary"
office2pdf document.docx --pdf-a
office2pdf report.docx --font-path /usr/share/fonts/custom

On macOS, office2pdf automatically searches Microsoft Office app fonts and local Office font caches before falling back to regular system fonts. --font-path is only needed as an override for custom local fonts.

WASM (Browser / Node.js)

Build with wasm-pack:

wasm-pack build crates/office2pdf --target web --features wasm

The default build does not add a CJK font. For zero-configuration Simplified Chinese fallback, opt in to the 3.3 MiB GB2312 subset (this feature implies wasm):

wasm-pack build crates/office2pdf --target web --features wasm-cjk-font

Use from JavaScript:

import init, {
  Office2PdfConverter,
  convertDocxToPdf,
  convertToPdf,
} from './pkg/office2pdf.js';

await init();

const docxBytes = new Uint8Array(await file.arrayBuffer());
const pdfBytes = convertDocxToPdf(docxBytes);

// Or use the generic API with a format string
const pdfBytes2 = convertToPdf(xlsxBytes, "xlsx");

// Register a font for this converter and make it the final fallback for every
// emitted font chain. The result-bearing API preserves structured warnings.
const fontBytes = new Uint8Array(await fontFile.arrayBuffer());
const converter = new Office2PdfConverter();
converter.registerFont(fontBytes);
converter.setLastResortFontFamily("Noto Sans SC");

const result = converter.convertDocxToPdf(docxBytes);
for (let index = 0; index < result.warningCount; index += 1) {
  const warning = result.warningAt(index);
  console.warn(warning.kind, warning.from, warning.to, warning.message);
}
const pdfBytes3 = result.pdf;

The compatibility functions return PDF bytes directly: convertToPdf(data, format), convertDocxToPdf(data), convertPptxToPdf(data), and convertXlsxToPdf(data). Their *WithResult counterparts return ConversionResult, which exposes pdf, warningCount, and warningAt(index). Office2PdfConverter provides the same result-bearing methods together with registerFont, clearFonts, setLastResortFontFamily, and clearLastResortFontFamily.

Browser and Node.js builds use the bundled Typst fallback fonts and also honor font faces embedded inside DOCX and PPTX files. Filesystem font paths remain unavailable in WASM, but callers can supply standalone TTF, OTF, or TTC bytes through Office2PdfConverter. When no registered or embedded face covers a CJK run, the result-bearing API emits a fallback-used warning whose to value is .notdef in the default build.

With wasm-cjk-font, Noto Sans CJK SC is automatically appended when the caller has not configured another last-resort family. The subset covers the complete GB2312 repertoire (7,445 characters, including 6,763 Han characters) plus printable ASCII. It is not full Traditional Chinese, Japanese, or Korean coverage. The warning to value identifies Noto Sans CJK SC when the bundle is used. An explicit setLastResortFontFamily call takes precedence.

Native Rust callers can use the same per-conversion path through ConvertOptions::font_bytes and ConvertOptions::last_resort_font_family.

CLI Options

Supported Formats

License

Licensed under Apache License, Version 2.0.