Document Redaction: clean documents before sharing
How BeeSensible's Document Redaction module works and what it produces.
Document Redaction lets your team open a document, see exactly what sensitive data it contains, and produce a clean version before sharing it or sending it to an AI tool.
What it does
- A user opens the Document Redaction tool in the desktop app, or the Documents section of the dashboard.
- They open or upload a PDF.
- BeeSensible scans the document and highlights sensitive data: names, IBANs, BSNs, addresses, and anything else in your active detection profile.
- The user chooses what to do with each finding: remove it, replace it with a placeholder, or leave it in.
- The tool produces a clean copy.
The original document is never modified. The clean copy is a separate file.
Redaction profiles
A redaction profile is a named, reusable set of the data types BeeSensible looks for in a document. Each profile simply turns entity types on or off, grouped by category (names, financial, health, and so on). Unlike the detection profiles in Realtime Privacy, document profiles have no per-app rules and no severity levels: they only decide what gets found.
BeeSensible ships a built-in profile to start from. It is marked BeeSensible and cannot be edited or deleted. From Documents > Profiles an admin can add their own with New profile, give it a name and description, and toggle exactly which types it detects.
One profile is the default, marked with a star. The default is applied automatically whenever anyone redacts a document, and it shows as the organisation default in the redaction toolbar. To change it, open Profiles and choose Set default on another profile.
When someone redacts a document, they can switch to another available profile from the toolbar for that document, so a stricter or lighter profile is one click away without changing the organisation default.
Where processing happens
Document Redaction can run on-device or in the EU cloud, depending on your organisation's setting:
- On-device. The desktop app processes the document entirely on the user's machine. The text does not leave the device.
- EU cloud. The document is sent to BeeSensible's servers inside the EU, processed in working memory, and discarded. Nothing is stored.
Both modes use the same detection and produce the same result. The active mode is shown at the bottom of the redaction panel.
Which engine runs
Document Redaction uses the same detection engines as Realtime Privacy: Bombus, the preferred engine, with Osmia as the automatic fallback. Both cover the same 65 data types. The engine is an organisation-level setting, not something you pick per document. See Detection engines for how Bombus and Osmia differ.
Supported document types
BeeSensible Document Redaction supports PDF files. Scanned PDFs without a text layer (no OCR) are not supported yet.
What the admin sees
Document Redaction activity appears in the admin dashboard as a count of documents processed and the types of sensitive data found. No document content is ever sent to the dashboard. Counts are aggregated per user and per document type.
Who can use it
Document Redaction is available to all team members who have the extension or desktop app installed. No special configuration is required beyond having the module enabled in your organisation's settings.