Open source · Runs locally · No cloud

De-identify medical documents. Locally.

Drop in a PDF or DOCX. Redakt finds and redacts personal data using a local AI model. Nothing leaves your machine. Ever.

Download Redakt View on GitHub

Free & open source

100% Local

All processing runs on your machine. No API keys, no cloud, no data transfer. HIPAA/KVKK/GDPR-compliant by design.

One click

Open a file. Click scan. Export. No terminal, no coding, no configuration. Works out of the box for clinicians and researchers.

Age conversion

Dates can be converted to patient age at time of visit. Preserves temporal information while removing identifiable dates.

See it in action

Drop a document. Click scan. Done.

REDAKT LOCAL DE-IDENTIFICATION
SCAN FOR PII EXPORT
NAME DATE ID LOCATION AGE
Document text
Redacted preview
Click “Scan” to preview redacted output.
READY 100% LOCAL
Human thinking belongs to humans. Mechanical work belongs to machines.
Read the Manifesto

Download Redakt

No installation wizards. No dependencies. Download, open, and start de-identifying.

MAC Apple Silicon
MAC Intel
WINDOWS x64
LINUX x64 (.deb)

The app is ~8 MB. Choose from 3 model sizes (6 GB to 21 GB) — downloads automatically on first launch.
macOS: Open the DMG and drag Redakt to Applications.
Windows: Run the installer (.exe) and follow the prompts.
Linux: Install with sudo dpkg -i Redakt_*.deb. RPM also available on GitHub.

Nothing downstream works without de-identification. We built that.