The LibreOffice PII Problem
LibreOffice is the most widely deployed open-source office suite, with an estimated 200 million users (The Document Foundation) and adoption across governments, universities, and enterprises that avoid Microsoft licensing costs. Yet until now, LibreOffice users had no integrated tool for detecting and anonymizing personally identifiable information (PII) within their documents.
The workaround — copy text into a web app, anonymize, paste back — destroys formatting, breaks document structure, and doubles the work. For Writer documents with complex styles, tables, and paragraph formatting, this copy-paste approach is unusable.
anonym.legal's LibreOffice Extension solves this by bringing the same 285+ entity type detection directly into Writer, Calc, and Impress.
Installation in Under 2 Minutes
The extension ships as a standard .oxt file — the universal LibreOffice extension format.
Quick Install:
- Download the .oxt file from anonym.legal/api/download/libreoffice
- Double-click the downloaded file
- LibreOffice opens and prompts you to confirm installation
- Restart LibreOffice
Manual Install:
- Open LibreOffice
- Go to Tools, then Extension Manager
- Click Add and select the .oxt file
- Restart LibreOffice
After restart, the anonym.legal sidebar appears under View, then Sidebar. Sign in with your anonym.legal account to start analyzing documents.
How Detection Works in LibreOffice
The extension uses the same hybrid detection engine as all anonym.legal products:
| Engine | Scope | Confidence |
|---|---|---|
| Regex patterns | Structured data (emails, SSNs, IBANs, credit cards, phone numbers) | 1.0 (deterministic) |
| spaCy NLP | Names, organizations, locations | 0.85+ (ML-based) |
Detection runs on the anonym.legal API — text is sent to ISO 27001-certified servers in Germany, analyzed, and results returned to the extension. No document files are uploaded; only extracted text is transmitted.
Per-application behavior:
- Writer: Full text extraction preserving paragraph boundaries, or selection-only mode
- Calc: Cell-based extraction as tab-separated values across all sheets
- Impress: Text object extraction from all slides
5 Anonymization Methods
The same 5 methods available across all anonym.legal products:
| Method | Example | Reversible | GDPR Classification |
|---|---|---|---|
| Replace | John Smith to PERSON | No | Anonymization (Recital 26) |
| Mask | john@example.com to j***@e***.com | No | Anonymization |
| Redact | 555-0123 to [REDACTED] | No | Anonymization |
| Hash (SHA-256/512) | SSN to a1b2c3d4... | No | Pseudonymization (WP29) |
| Encrypt (AES-256-GCM) | Name to ENC:xyz... | Yes | Pseudonymization (Art. 4(5)) |
Per-entity operator selection lets you apply different methods to different PII types in a single pass — for example, replace names but encrypt email addresses.
Format Preservation in Writer
The critical advantage over copy-paste workflows: the extension preserves document formatting during anonymization.
7 font properties preserved:
- Font family, size, weight (bold), style (italic), color, underline, strikethrough
4 paragraph properties preserved:
- Alignment, first-line indent, left margin, right margin
When "John Smith" appears in bold 14pt Times New Roman, the replacement "PERSON" appears in bold 14pt Times New Roman. Styles, tables, headers, footers — all remain intact because the extension operates through LibreOffice's UNO API document model rather than text export/reimport.
Note: Format preservation is Writer-only. Calc and Impress perform text-level replacement without format property tracking.
Preview Before You Apply
The preview dialog shows up to 50 detected entities in a table:
- Entity type (PERSON, EMAIL_ADDRESS, PHONE_NUMBER, etc.)
- Original text
- Proposed replacement
- Confidence score
You can:
- Accept or reject each entity individually
- Select all / deselect all
- Review per-entity before any changes are applied
Only accepted entities are anonymized when you click Apply. Ctrl+Z undoes the last operation in Writer.
Presets and Cross-Device Sync
Save anonymization configurations as presets — which entity types to detect, which methods to apply, confidence thresholds — and reuse them across documents.
On paid plans, presets sync automatically across all anonym.legal platforms:
- Set up a preset in the web app
- It appears in LibreOffice, the Desktop App, and the Office Add-in within 5 minutes
- Encryption keys sync with ZK wrapping (XChaCha20-Poly1305) — only your password can decrypt them
LibreOffice vs. Microsoft Office: Same Engine, Same Results
| Feature | LibreOffice Extension | Office Add-in |
|---|---|---|
| Entity types | 285+ | 285+ |
| Languages | 48 | 48 |
| Methods | 5 (Replace, Mask, Redact, Hash, Encrypt) | 5 (Replace, Mask, Redact, Hash, Encrypt) |
| Format preservation | Writer only | Word only |
| Preview dialog | Up to 50 entities | Up to 50 entities |
| ZK authentication | Yes | Yes |
| Preset sync | Yes (paid plans) | Yes (paid plans) |
| Platforms | Windows, macOS, Linux | Windows, Mac, Web, iPad |
Both extensions use the same backend detection engine. A document anonymized in Word produces the same entity detection results as the same text processed in Writer.
System Requirements
- LibreOffice 3.0 or later
- Python 3 (bundled with most LibreOffice installations)
- Internet connection (for PII detection API)
- anonym.legal account (free tier: 200 tokens/month)
Getting Started
- Download the extension (.oxt file)
- Double-click to install, restart LibreOffice
- Open View, then Sidebar, then select the anonym.legal panel
- Sign in with your account
- Open a document, click Analyze, review results, click Apply
Full documentation: LibreOffice Extension Docs
Sources:
- The Document Foundation — estimated 200M+ LibreOffice users worldwide
- anonym.legal PII Detection Testing — 95.5% accuracy, 42/44 independent tests
- GDPR Recital 26 — anonymization removes data from GDPR scope; Article 4(5) — pseudonymization definition