PIIShield uses the following open-source software components. We gratefully acknowledge their authors and contributors.
Cross-platform desktop application framework.
Generalist model for named entity recognition (NER). Used for detecting person names, organizations, addresses, and other entities.
High-performance inference engine for machine learning models.
Optical character recognition engine for extracting text from images and scanned documents.
Pure-Python library for PDF text extraction and metadata inspection.
Python wrapper around pdftoppm (Poppler) for high-fidelity PDF page rendering and image conversion.
Python Imaging Library for image drawing, pixel manipulation, and PDF document reconstruction.
Library for reading and writing Microsoft Word (.docx) files.
Library for reading and writing Excel (.xlsx) files.
Computer vision library used for image preprocessing, rotation, and manipulation.
Library for generating realistic fake data. Used for pseudonymization of detected PII.
Display heading font. Designed by Tokotype.
Monospace font for code and technical data. Designed by JetBrains.
Icon font by Google. Used for UI iconography throughout the application.