LocalRedact

LocalRedact is a free PDF redaction tool for people who want documents processed on their own machine and can work with Python, a command line, or an API. Its local workflow and output checks are useful safeguards, but its alpha status and fallible detection make human review essential for consequential documents.

Overview

LocalRedact reads PDF text and coordinates locally through PyMuPDF, without an LLM, cloud API, telemetry, or model download. Redactions remove underlying text and blank pixels where they intersect images; default sanitization also removes common extras such as metadata, embedded files, JavaScript, links, and saved form state. That is a strong privacy posture for sensitive files, though the PDF-only input and lack of batch processing limit its fit for broader document operations.

After producing a file, LocalRedact reopens it and checks that values it detected are no longer extractable. This is more meaningful than simply drawing black boxes over text, but it only checks detected values: it cannot establish that every sensitive item was found. The project calls its detection best effort, not compliance certification, and warns that unusual formatting, handwriting, weak OCR, vector-outline text, novel identifiers, and sensitive prose may escape it.

Key features

Rule-based targeting

Plain-English instructions can target personal, financial, government, health, and credential data, along with quoted exact values and selected page or layout conditions. Deterministic rules give users a local alternative to sending documents to an AI service. They do not remove the need to inspect results, especially where content falls outside those targets or is formatted unusually.

OCR and interfaces

Scanned PDFs can use locally installed Tesseract through an OCR option. Image-only pages fail closed without OCR by default, which avoids silently treating unreadable pages as safe, but successful OCR still does not guarantee detection. The package provides a Python API and two command-line tools; an optional HTTP service is installable with the server extra. That service binds to 127.0.0.1 by default and makes no outbound requests, while non-loopback binding requires an explicit option.

Pricing

LocalRedact costs nothing under the LocalRedact plan, an AGPL-3.0-only Python package with an optional local HTTP service. It is distributed through PyPI and requires Python 3.10 or later. There is no free trial because the software is free, and no paid tier is described. This suits users comfortable running Python software who want local processing without a subscription; it is not a managed service for teams seeking a supported, certified workflow.

The license terms matter for adoption: LocalRedact uses GNU AGPL version 3 only, and PyMuPDF has separate AGPL and commercial licensing. Organizations should account for both before incorporating the software into a deployment.

Platforms

LocalRedact supports API use, Linux, macOS, Windows, and self-hosted deployment. The desktop deployment label does not imply a graphical desktop application: the stated interfaces are a Python API, command-line tools, and optional loopback HTTP service. Its local design is suitable where documents should stay on the user's system, but it accepts PDFs only and does not provide batch processing.

Who it's for

LocalRedact fits individuals and technical teams handling sensitive PDFs who value local processing and can incorporate a Python package, CLI, or API into their workflow. Its output reopening and extractability check add a useful verification step, and local Tesseract support extends the workflow to scans. It is a poor choice for unattended high-stakes redaction, compliance-dependent work, or workflows that need batch processing: the software is alpha, detection can miss content, and the project explicitly disclaims compliance certification.

Pros and cons

• Pro: Local processing avoids sending documents to an LLM or cloud API, a practical advantage for privacy-sensitive PDFs.
• Pro: Redactions remove underlying text as well as covering intersecting image pixels, and default sanitization strips common PDF extras.
• Pro: Reopening the output to check detected values improves assurance beyond visual masking alone.
• Con: Best-effort detection can miss several classes of content, so high-stakes outputs need visual review.
• Con: Alpha status and the lack of compliance certification make it unsuitable as a standalone control for regulated workflows.
• Con: PDF-only input and no batch processing restrict throughput and format flexibility.
• Con: Python 3.10 or later and AGPL licensing require technical and licensing fit; PyMuPDF's separate licensing also needs consideration.

Alternatives

For another free, locally oriented option, BleachPDF runs on Linux and macOS on the user's own computer and has no subscriptions; LocalRedact is the better fit when its stated API, Windows support, OCR option, or detection and output checks matter. Choose PDFRedaction if a web platform and stated daily allowance of 10 documents for registered free users, AI detection, OCR, image redaction, and local processing suit the workflow better. For a macOS or Windows desktop tool, PineThink Redact offers single-document personal use without an account on its free plan, with a 14-day free trial for Premium.

Philter is another option for API, Linux, self-hosted, and Windows use; its free desktop plan is for individual, non-commercial, and evaluation use and includes on-device redaction. RedactFile may suit users who prefer web access or an API and can work within its free plan's 2MB upload size, 10 daily uploads, and 500MB storage. RedactLaw offers web, API, and self-hosted access, with a free allowance of five documents per month and up to 10 pages per document. Redactable is a web alternative with three free documents total before its pay-as-you-go option. Excise is another free option available through API, Linux, macOS, and Windows.

Browse more options in Document Redaction Software.

Verdict

Choose LocalRedact if you need free, local PDF redaction with rule-based targeting, output checks, and Python, command-line, or API access. Its privacy-conscious processing and sanitization make it a credible fit for technical users working with sensitive files. Look elsewhere if you need batch processing or compliance certification, and visually review every high-stakes output because the software is alpha and detection can miss content.