dxpdf

A fast, lightweight DOCX-to-PDF converter written in Rust, powered by Skia.

Convert Microsoft Word .docx files to high-fidelity PDF documents — from the command line or as a Rust library. No Microsoft Office, LibreOffice, or cloud API required.

Built by nerdy.pro.

Why dxpdf?

• Fast — converts a 24-page document in ~115ms on Apple Silicon
• Accurate — Flutter-inspired measure→layout→paint pipeline with pixel-level fidelity
• Standalone — no external dependencies beyond Skia; no Office installation needed
• Cross-platform — runs on macOS, Linux, and Windows
• Dual-use — works as a CLI tool, Rust library (use dxpdf;), or Python package (import dxpdf)

Quick Start

Install

cargo install dxpdf

Convert a file

dxpdf input.docx                  # outputs input.pdf
dxpdf input.docx -o output.pdf    # specify output path

Use as a library

let docx_bytes = std::fs::read("document.docx")?;
let pdf_bytes = dxpdf::convert(&docx_bytes)?;
std::fs::write("output.pdf", &pdf_bytes)?;

You can also inspect or transform the parsed document model:

use dxpdf::{parse, model};

let document = parse::parse(&std::fs::read("document.docx")?)?;

for block in &document.blocks {
    match block {
        model::Block::Paragraph(p) => { /* ... */ }
        model::Block::Table(t) => { /* ... */ }
    }
}

let pdf_bytes = dxpdf::convert_document(&document)?;

Use from Python

pip install dxpdf

import dxpdf

# Bytes in, bytes out
pdf_bytes = dxpdf.convert(open("input.docx", "rb").read())

# File to file
dxpdf.convert_file("input.docx", "output.pdf")

What's Supported

dxpdf handles the most common DOCX features used in real-world business documents:

Building from Source

Prerequisites

• Rust toolchain (1.70+)
• clang (required by skia-safe for building Skia bindings)

Build

cargo build --release

The release binary will be at target/release/dxpdf.

Architecture

dxpdf follows a measure→layout→paint pipeline inspired by Flutter's rendering model:

DOCX (ZIP) → Parse → Document Model (ADT) → Measure → Layout → Paint → Skia PDF

Each layout element (paragraphs, table cells, headers/footers) goes through three phases:

1. Measure — collect fragments, fit lines, produce draw commands with relative coordinates
2. Layout — assign absolute positions, handle page breaks, distribute heights (e.g., vMerge spans)
3. Paint — emit draw commands at final positions (shading → content → borders)

Modules

Running Tests

cargo test

The test suite includes 104 unit tests and 9 integration tests covering layout, tables, lists, floats, headers/footers, hyperlinks, superscript/subscript, field codes, and end-to-end conversion.

Visual regression tests compare rendered PDFs against Word-generated references using pixel matching (see VISUAL_COMPARISON.md).

OOXML Feature Coverage

Validated against ISO 29500 (Office Open XML). 34 features fully implemented, 6 partial, 15 not implemented.

Text Formatting (w:rPr)

Paragraph Properties (w:pPr)

Styles

Tables

Images

Page Layout

Headers/Footers

Lists

Fields

Other

Performance

Benchmarked on Apple M3 Max with hyperfine (20 runs, 3 warmup):

See BENCHMARKS.md for full history.

Dependencies

Contributing

Contributions are welcome. Please open an issue before submitting large PRs.

Built by nerdy.pro.

License

MIT