pdfmasker

With this library you can mask sensitive text in PDFs or basically change any text in your PDF files.

Usage

from pdfmasker import mask_pdf

pdf_bytes = open("paystub.pdf", "rb").read()

result = mask_pdf(pdf_bytes, patterns=["Jane Doe", "123-45-6789"])

result.document_content   # bytes — the masked PDF
result.entries            # per-target log: target, replacement, kind, count

By default each match is replaced in place with a run of X. An editor chooses how matches are rendered — a constant string, or consistent pseudonyms:

from pdfmasker import FixedStringEditor, mask_pdf

mask_pdf(pdf_bytes, patterns=["Jane Doe"], editor=FixedStringEditor("[REDACTED]"))

Pluggable parts

Masking is composed from four pluggable parts:

• Backend — how the PDF is read and rewritten. The default in-process PikepdfBackend edits content streams directly in Python.
• Detector (optional) — what to mask. Pass literal patterns, or configure detectors such as RegexDetector to discover targets from the document text.
• Editor — how each match is rendered: FixedCharEditor (default X), FixedStringEditor, PseudonymizeEditor, or KeyedPseudonymizeEditor.
• SubstitutionStore — keeps pseudonyms consistent and collision-free. Share one across documents so the same value always masks to the same replacement.

from pdfmasker import InMemorySubstitutionStore, Masker, PseudonymizeEditor, RegexDetector

store = InMemorySubstitutionStore()
masker = Masker(
    detectors=[RegexDetector(r"\d{3}-\d{2}-\d{4}", kind="ssn")],
    editor=PseudonymizeEditor(store),
)
result = masker.mask(pdf_bytes, patterns=["Jane Doe"])

Contributing

The architecture, repository layout, setup, dev commands (make help), the build/cross-build model, and the release flow all live in CONTRIBUTING.md.