5× faster than industry leaders

The Fastest PDF Library

For Rust, Python, JavaScript, TypeScript, Golang, C# and WASM

0.8ms

Mean per PDF

100%

Pass rate

3,830

PDFs tested

5.1×

Faster than #2

Get started View on GitHub

Up and running in seconds.

One install. Three lines of code.

$ cargo add pdf_oxide

use pdf_oxide::PdfDocument;

let mut doc = PdfDocument::open("paper.pdf")?;
let text = doc.extract_text(0)?;
let images = doc.extract_images(0)?;

One Library.
Create|Edit|Extract

Every PDF operation in a single dependency.
No wrappers, no subprocess calls, no C/C++/Java runtimes.

Create

Build PDFs from any source format.

Markdown
Convert Markdown to pixel-perfect PDFs with headings, lists, tables, and code blocks.
HTML
Turn HTML markup into structured PDF documents with full CSS layout support.
Images
Single or multi-page PDFs from PNG, JPEG, and TIFF with automatic sizing.
QR & Barcodes
Code128, EAN-13, UPC-A, and QR codes with configurable error correction.
Builder API
Fluent PdfBuilder chain for page size, margins, fonts, metadata, and headers.
Forms
Text fields, checkboxes, radio buttons, dropdowns, stamps, and watermarks.

Edit

Modify any part of an existing PDF.

DOM Editing
Find text, replace content, and restyle — navigate the PDF like a web page.
Pages
Rotate, crop, merge documents, extract page ranges, and reorder.
Forms
Get and set field values, add or remove fields, flatten to static content.
Annotations
Add highlights, notes, and links. Modify or flatten selectively.
Images
Reposition, resize, and replace embedded images with exact bounds.
Security
AES-256 encryption, passwords, and fine-grained permission flags.

Extract

Get everything out of any PDF.

Text
Full-page text, styled spans with font metadata, or per-character positions.
Images
Content streams, nested Form XObjects, and inline images with color spaces.
Markdown
Clean Markdown or HTML with heading detection and table preservation.
Forms
All field values and types. Export to FDF or XFDF. XFA analysis.
Metadata
XMP, Dublin Core, page labels, catalog, and trailer dictionaries.
Search
Full-text search with regex, case-insensitive, and whole-word modes.

5× faster than every alternative.

Benchmarked on 3,830 real-world PDFs from 3 public test suites.

Node.js, Go, and C# share the same Rust core — expect matching numbers.

Library	Language	Mean	p99	Pass Rate	License
PDF Oxide	7 languages	0.8ms	9ms	100%	MIT
PyMuPDF	Python	4.6ms	28ms	99.3%	AGPL-3.0
oxidize_pdf	Rust	13.5ms	11ms	99.1%	MIT
pypdfium2	Python	4.1ms	42ms	99.2%	Apache-2.0
pdfminer	Python	16.8ms	124ms	98.8%	MIT
pdfplumber	Python	23.2ms	189ms	98.8%	MIT
pypdf	Python	12.1ms	97ms	98.4%	BSD-3
unpdf	Rust	2.8ms	10ms	95.1%	MIT
pdf_extract	Rust	4.08ms	37ms	91.5%	Apache-2.0
lopdf	Rust	0.3ms	2ms	80.2%	MIT

Measured on 3,830 PDFs (veraPDF, Mozilla pdf.js, DARPA SafeDocs).
Single-thread, no warm-up, 60s timeout.

100% reliable. Zero surprises.

Tested on 3,830 PDFs from three independent public test suites.
Zero panics. Zero timeouts. Zero crashes.

3,823

valid PDFs parsed
without a single failure

99.5%

text quality parity
vs PyMuPDF & pypdfium2

intentionally broken PDFs
correctly rejected

The corpus covers every PDF version (1.0–2.0), encrypted files, malformed documents, CJK encodings, and DARPA SafeDocs security edge cases designed to crash vulnerable parsers.

Start building.

Free for commercial and open-source projects.

Read the docs