Open Source Licenses

Built on open source

DukPdf's local-first processing is only possible because of exceptional open-source projects. This page lists the components we bundle or load, their licenses, and where to find their source code. Thank you to every maintainer and contributor behind them.

PDF and document engines

These do the heavy lifting — entirely on your device — in the web app, the desktop app, or both.

qpdf — Apache License 2.0. PDF page operations (merge, split, rotate, encryption). Native binary on desktop; qpdf-wasm (Apache-2.0) in the browser. Source

PDFium — BSD 3-Clause License. PDF rendering and rasterization (thumbnails, previews, redaction burn-in) on desktop, via the pdfium-render bindings (MIT/Apache-2.0). Source

pdf.js — Apache License 2.0. In-browser PDF rendering and text extraction. Served from our own origin. Source

pdf-lib — MIT License (with @pdf-lib/fontkit, MIT). In-browser PDF creation and editing. Source

LibreOffice — Mozilla Public License 2.0 (portions LGPL-3.0-or-later). Word/Excel/PowerPoint → PDF conversion on desktop. We ship the official, unmodified build from The Document Foundation — on macOS it downloads on first use (from our mirror, byte-for-byte identical to the official archive), on Windows it is bundled. Copyright © The Document Foundation and contributors. Licenses · Source code

ZetaOffice (zetajs) — Mozilla Public License 2.0. LibreOffice compiled to WebAssembly by allotropia, powering Word/Excel/PowerPoint → PDF in the browser. Unmodified build, self-hosted on our CDN. Source

OCR and machine learning

Tesseract — Apache License 2.0. OCR engine: native on desktop, tesseract.js (Apache-2.0) in the browser, with language models from the tessdata project. Source

PaddleOCR PP-OCR models — Apache License 2.0. Neural OCR model weights used by the secondary OCR path. Source

ONNX Runtime — MIT License. ML inference (desktop native and onnxruntime-web in the browser), with the ort Rust wrapper (MIT). Source

Transformers.js — Apache License 2.0. Browser-side model tooling from Hugging Face. Source

Application framework

Tauri — MIT / Apache License 2.0 (dual). Desktop app runtime and plugins (dialog, fs, store, updater, deep-link, shell, and more). Source

Next.js & React — MIT License. The web application framework. Next.js · React

Supabase libraries — MIT License. Authentication and API clients (supabase-js, ssr). Source

lopdf — MIT License. Low-level PDF object manipulation in the desktop engine. Source

Rust crate ecosystem— MIT and/or Apache-2.0. Including image, printpdf, rayon, serde, flate2, zip, reqwest, keyring, and others — the full list with versions is in the app's Cargo.toml manifest.

JavaScript ecosystem— MIT and similar. Including Tailwind CSS, dnd-kit, react-window, jszip, docx, and others — the full list with versions is in the app's package.json manifest.

Fonts

The DukPdf interface uses Nunito and Open Sans, both under the SIL Open Font License 1.1. Fonts offered inside tools (for watermarks, page numbers, and text) are open-licensed fonts fetched from Google Fonts under their respective licenses and embedded into the PDFs you create.

Full license texts and source availability

Each component above keeps its own license; the links go to the canonical upstream source where the complete license text lives. We redistribute LibreOffice and ZetaOffice unmodified — their source code is available from the upstream projects linked above, as the Mozilla Public License requires.

If you believe an attribution is missing or incorrect, or you'd like a copy of any notice, email support@dukpdf.com and we'll fix it promptly.