Every PDF chore, handled on your own machine. MyPDF is a small, fast desktop app for Windows that covers the everyday PDF workflow you would normally take to an online converter, except nothing ever leaves your laptop. No uploads, no accounts, no limits, no tracking. Built for people who handle sensitive documents and would rather not send them to someone else's server.
| 🧩 Merge | Stack several PDFs into one, in any order |
| 🗂️ Arrange | See every page as a thumbnail. Drag to reorder, rotate, or remove pages |
| ✂️ Split | Break a PDF apart per page or by custom ranges |
| 📑 Pick Pages | Keep only the pages you need |
| 🗜️ Compress | Presets, full manual control, or "fit a size": tell it the MB limit and it finds settings that fit |
| 🔄 Rotate | The whole file or selected pages |
| 🖼️ PDF to Images | Render pages as PNG or JPG |
| 📄 Images to PDF | Photos and scans into one tidy PDF |
| 📝 Office to PDF | Word, Excel, PowerPoint via LibreOffice |
| ✏️ PDF to Word | Editable .docx out of any PDF |
| 🔒 Lock / Unlock | AES password protection, or remove it from your own files |
| 💧 Watermark | Faint text across every page |
| 🔍 OCR | Make scans searchable, with live per page progress |
| 🔤 Extract Text | All the text into a .txt file |
| 🔢 Page Numbers | Neat page numbers, covers skippable |
| 🏷️ Metadata | Edit title, author, subject and keywords |
Most tools accept many files at once and process them as a parallel batch, with a Cancel button for long jobs. The interface speaks English and Bahasa Indonesia. Plus the small things that make it pleasant: drag and drop anywhere, page thumbnails with file info, live progress bars, a recent work history, output files that never overwrite existing ones, undo and keyboard control in the page organizer, Esc to jump back home, and a settings page for your name, language, default OCR language and output folder.
python tests/engine_smoke.py # engine test suite, core deps only
npm run build # type check + frontend buildBoth run automatically on every push via GitHub Actions.
React UI → thin Tauri (Rust) command → Python engine (stdin/stdout JSON)
The interface is React + TypeScript in a Tauri shell. All PDF logic lives in a single Python engine (engine/pdf_engine.py) built on PyMuPDF and pikepdf, which streams JSON progress lines so the UI can show real percentages. Heavier features call battle tested externals per task: Ghostscript for compression, LibreOffice for Office conversion, and Tesseract + ocrmypdf for OCR.
Grab the latest setup exe from the releases page and run it.
Windows SmartScreen note. The installer is not code signed (signing certificates cost money this free project does not have), so Windows may show "Windows protected your PC". Click More info, then Run anyway. The entire source code is in this repository and every release is built from it in public by GitHub Actions, so you can verify exactly what you are running.
The Windows installer runs out of the box. It bundles a Python runtime with all engine libraries (PyMuPDF, pikepdf, Pillow, pdf2docx, and the ocrmypdf package), so no Python or pip setup is needed.
Three features need free external programs that are not bundled with the installer. The app detects what is installed and tells you exactly what is missing before you run anything:
On Linux, macOS, or when running from source, you need Python 3.10+ with
pip install pikepdf pymupdf pillow (plus pdf2docx and ocrmypdf for those
features).
Optional, per feature (the app detects what is installed and tells you):
| Feature | Tool | Notes |
|---|---|---|
| Strong compression | Ghostscript | Without it, compression still works but stays mild |
| Office to PDF | LibreOffice | Required for this tool |
| OCR | Tesseract | The OCR engine package ships with the installer; only the Tesseract program is an extra install |
| PDF to Word | pip install pdf2docx |
Only when running from source; the Windows installer already includes it |
For OCR in languages beyond English, place the matching .traineddata from
tessdata_fast into your
tessdata folder (set TESSDATA_PREFIX if you keep it outside the Tesseract
install directory).
Prerequisites: Node.js, Rust (MSVC toolchain on Windows), Python 3.10+.
npm install
npm run tauri dev # run in development
npm run tauri build # produce the installer (MSI + NSIS)The engine also works standalone, which makes it easy to test:
echo '{"task":"doctor","params":{}}' | python engine/pdf_engine.pyEverything runs locally. The app makes no network requests: fonts are bundled, processing is on device, and your files stay in your folders.
AGPL 3.0. The engine builds on PyMuPDF (AGPL licensed), so this project shares the same license. In short: use it, modify it, ship it, but keep the source open. The full source for every release is published in this repository, and each release is built from it in public by GitHub Actions.
Bundled and optional components keep their own licenses: pikepdf, Pillow, pdf2docx, and ocrmypdf (Python packages inside the Windows installer), plus Ghostscript, LibreOffice, and Tesseract when installed separately as per feature helpers.
Issues and pull requests are welcome. Adding a new tool is usually one Python
function in engine/pdf_engine.py plus one entry in the tool list in
src/App.tsx.