LocalRedact
github.com
- Free plan
- Windows
- macOS
- Linux
- API
- Self-hosted
What LocalRedact is
LocalRedact is a PDF redaction tool that turns plain-English instructions into deterministic rules. It can target personal, financial, government, health, and credential data, along with quoted exact values and selected page or layout instructions. Redactions remove underlying text and blank image pixels where they overlap. Processing happens locally: the project says it uses no LLM, cloud API, telemetry, or model download, and reads PDF text and coordinates with PyMuPDF. By default, it removes common PDF extras such as metadata, embedded files, JavaScript, links, and saved form state, then reopens the output to check whether detected values remain extractable. It offers a Python API, two command-line tools, and an optional HTTP service that binds to 127.0.0.1 by default. Scanned PDFs can use locally installed Tesseract; image-only pages fail closed without OCR by default. LocalRedact is free and open source under GNU AGPL version 3 only. The project calls it alpha software, describes detection as best effort rather than compliance certification, and recommends visual review of high-stakes output. It requires Python 3.10 or later.
Who it suits
LocalRedact suits people who need to redact PDFs locally through an API, command line, or optional local HTTP service. Its alpha status and stated detection limits make visual review important, especially for high-stakes documents.
Good to know, and what to watch
- Processes PDF text and coordinates locally
- Removes underlying text and intersecting image pixels
- Checks output for extractable detected values
- Offers Python, command-line, and optional HTTP interfaces
- Alpha software with best-effort detection
- May miss handwriting, unusual formatting, or sensitive prose
- Not a compliance certification
- Batch processing is not supported
Plans
Every plan and priceLocalRedact
Open source
Getting started
- Open the project at https://github.com/1aifanatic/localredact
- Install the Python package from PyPI with Python 3.10 or later
- Use the Python API or either command-line tool with a PDF
- Install the optional server extra if you want the loopback HTTP service
- Install Tesseract locally and enable the OCR option for scanned PDFs
- Visually review high-stakes redacted output before sharing
Limits to check
LocalRedact describes detection as best effort and warns that unusual formatting, handwriting, weak OCR, vector-outline text, novel identifiers, and sensitive prose may be missed. Image-only pages fail closed without OCR, and the project recommends visual review for high-stakes files.
The verdict
LocalRedact offers local PDF redaction with multiple interfaces and checks on saved output. Its stated detection limits mean the project recommends visually reviewing high-stakes results before sharing them.
Questions people ask
Is LocalRedact free and open source?
Yes. It is listed as free and open source, with a free plan. The license is GNU AGPL version 3 only.
What files does it support?
PDF is the listed input format. Scanned PDFs can use OCR through locally installed Tesseract.
Does processing leave the computer?
The project says processing is local and uses no LLM, cloud API, telemetry, or model download. Its local HTTP service binds to 127.0.0.1 by default and makes no outbound requests.
Which operating systems are supported?
LocalRedact is listed for Windows, macOS, and Linux. The package is distributed through PyPI and requires Python 3.10 or later.
Is it suitable for compliance certification?
No. The project describes detection as best effort and says it is not a compliance certification. It recommends visually reviewing every high-stakes output before sharing.
What the maker publishes
- Availability
- The package is distributed through PyPI and requires Python 3.10 or later. source
- Compliance
- The project describes its detection as best effort and says it is not a compliance certification. source
- Detection
- Supported targets include personal, financial, government, health, and credential data, plus quoted exact values and selected page or layout instructions. source
- Detection limits
- It says detection may miss unusual formatting, handwriting, weak OCR, vector-outline text, novel identifiers, or sensitive prose. source
- Integrations
- The base package depends on PyMuPDF, and the optional HTTP service can be installed with the server extra. source
- Interfaces
- LocalRedact provides a Python API, two command-line tools, and an optional loopback HTTP service. source
- Known limit
- The project describes detection as best effort and says it is not a compliance certification; it warns that unusual formatting, handwriting, weak OCR, and other content may be missed. source
- License
- LocalRedact is licensed under GNU AGPL version 3 only, and the README notes that PyMuPDF has separate AGPL and commercial licensing. source
- Limit
- The project describes itself as alpha software and warns automated redaction may miss sensitive content, recommending visual review of high-stakes output. source
- OCR
- Scanned PDFs can use locally installed Tesseract through an OCR option, while image-only pages fail closed without OCR by default. source
- Output checks
- It sanitizes common PDF extras by default, reopens the output, and checks that detected values are no longer extractable. source
- Privacy
- The project says it uses no LLM, cloud API, telemetry, or model download, and reads PDF text and coordinates locally with PyMuPDF. source
- Processing
- The project says processing is local and does not use an LLM, cloud API, telemetry, or model download. source
- Purpose
- LocalRedact is a private PDF redaction tool that processes plain-English instructions with deterministic rules. source
- Redaction
- It applies PDF redactions that remove underlying text and blank intersecting image pixels. source
- Review guidance
- The README calls the software alpha and recommends visually reviewing every high-stakes output before sharing it. source
- Sanitization
- By default, it removes common PDF extras including metadata, embedded files, JavaScript, links, and saved form state. source
- Service security
- The local HTTP service binds to 127.0.0.1 by default, makes no outbound requests, and requires an explicit option for non-loopback binding. source
- Verification
- After writing output, it reopens the PDF and checks that detected values are no longer extractable. source
Checked against the maker's pages on 8 Oct 2026.