PII Detection & Anonymization Features
PII Detection
Detect 320+ entity types across 70+ countries. 317 custom regex recognizers handle structured data (IDs, tax numbers, passports), while spaCy, Stanza, and XLM-RoBERTa NLP detect names, locations, and organizations. All models run on our own servers in Germany — zero data sent to third parties.
- 320+ entity types, 70+ countries
- 317 regex recognizers + NLP detection
- Confidence scoring with context analysis
Anonymization Methods
Choose from 7 proven anonymization methods: Replace, Redact, Hash, Encrypt, Asymmetric Encrypt, Mask, or Keep to fit your compliance requirements.
- Replace with fake data
- Redact completely
- Hash with SHA-256
Reversible Encryption
The only anonymization method that can be reversed. Encrypt PII with AES-256-GCM and restore original data anytime with your key.
- AES-256-GCM encryption
- Full data recovery
- Your key, your control
Multi-Language Support
Full support for 48 languages with intelligent NLP processing. Includes RTL support for Arabic, Hebrew, Persian, and Urdu.
- 48 languages supported
- 4 RTL languages (Arabic, Hebrew, Persian, Urdu)
- Language auto-detection
MCP Server
Native Claude Desktop integration plus HTTP for Cursor & VS Code. 10 privacy tools including batch analysis and image redaction.
- 10 privacy tools (text, batch, image)
- Batch analyze 1-100 texts in parallel
- Image OCR detection & redaction
Desktop App
Keep documents on your device while using cloud-powered entity detection. Only extracted text is sent for analysis.
- Documents stay on your device
- Military-grade vault encryption
- Encrypted local history
Office Add-in
Anonymize documents directly in Microsoft Word, Excel, and PowerPoint. Native integration with format preservation—work where you already are.
- Word, Excel & PowerPoint support
- Real-time PII detection
- One-click anonymization
Chrome Extension
Automatically anonymize PII before sending to ChatGPT, Claude, Gemini, DeepSeek, Perplexity, and Abacus.ai. Real-time protection with reversible encryption.
- ChatGPT, Claude, Gemini, DeepSeek, Perplexity, Abacus.ai
- JIT interception before send
- Auto de-anonymization of responses
Nextcloud App
Detect and anonymize PII directly in Nextcloud. Sidebar integration, right-click file action, and all 7 anonymization methods without leaving your file manager.
- Native Nextcloud 28–31 integration
- Right-click file anonymization
- Sidebar panel with live preview
Cloud Storage Addins
Anonymize files directly in Microsoft 365, Google Drive, Dropbox, and Nextcloud. Browse, analyze, and save results back to the cloud.
- Microsoft 365 OneDrive & SharePoint
- Google Drive, Dropbox & Nextcloud
- Save anonymized files back to cloud
API Integration
RESTful API for developers to integrate PII detection and anonymization into any application or workflow. Full documentation available.
- RESTful endpoints
- JWT authentication
- Rate limiting
Batch Processing
Process multiple documents at once with our batch processing feature. Enterprise plans support up to 100 documents per batch.
- Multi-document upload
- Parallel processing
- Progress tracking
Zero-Knowledge Security
Your password NEVER leaves your device. We use cryptographic proofs to verify you without ever seeing your password - the most secure way to protect your account.
- Password never transmitted
- 24-word recovery phrase
- Argon2id + XChaCha20-Poly1305
Image Redaction
Detect and redact PII in scanned documents, photos, and screenshots. OCR-powered with 38 language support.
- 38 OCR languages
- Tesseract-powered detection
- Bounding-box redaction
Why Choose cloak.business?
Regex-first architecture: structured data is fully deterministic, names and locations use transparent NLP with confidence scores
Our Approach: Regex + NLP
- Structured data: 100% reproducible (317 regex recognizers)
- Names & locations: NLP with confidence scores
- Every detection traceable to a specific pattern or model
- 48 languages across 3 NLP engines
- Fast, CPU-only performance
AI-Only Solutions
- All detections are probabilistic
- Can't explain why something was flagged
- Requires custom training datasets
- Model drift degrades accuracy
- GPU required, higher costs