Redact a PDF with Cloak
Review detected identifiers and choose the appropriate local or scan-processing flow.
Source reviewed Product documentation
Open the BeforeMay redaction tool for the public PDF workflow. Cloak detects possible identifiers, lets you review them and produces a redacted copy. Detection suggestions need your review before the file is shared.
Digital PDFs
The digital-PDF flow reads, detects and redacts in the browser. Drop the file, inspect the detected identifiers and pages, correct the selection and download the redacted copy. Keep the original separately as the source record.
Check that every identifier you intend to remove is covered. A detector can miss a value or mark an unrelated value. Inspect the downloaded copy, including pages with annotations and images, before giving it to another person.
Scans
A scan has no readable text layer for browser detection. The tool explains this and offers its scan-processing route where available. That route uses the Australian OCR service; it does not have the same no-upload boundary as the digital-PDF flow. Read the notice before submitting a scan.
Redaction and substitution
Use true redaction when the recipient must not recover the selected information. A demo copy with substitute values is pseudonymisation, not anonymisation. It can retain names, addresses and other information that identifies someone. See AI processing for the separate outbound gate used by Workpapers; manually redacting in Cloak is not a prerequisite that replaces that gate.
Developer access
Cloak's developer dashboard and redaction API use their own account and service. They are separate from the Workpapers MCP interface. Confirm the enabled endpoint, credentials and processing agreement with support before integrating; this guide does not assume that a browser-local conversion has an equivalent public API.