PII Detection: 317 Entity Types
PII detection is the process of finding personal data in text before it can leak. Our detection engine combines 317 custom regex recognizers with NLP models. Together, they identify 317 types of personal information across 70+ countries. Same input, same output - every time.
How Detection Works
Regex Pattern Matching (Structured PII)
317 custom PatternRecognizers use regex patterns to detect structured data. This includes national IDs, tax numbers, passports, and driver licenses. Each pattern uses boundary checks. This stops false matches in code or structured data.
NLP Named Entity Recognition (Names & Locations)
Three NLP models catch PII regex alone misses: spaCy (25 languages), Stanza (7 languages), and XLM-RoBERTa (16 languages). They spot names, places, and company names in plain text. Every model runs on our own servers in Germany. We never send your data to Meta, Google, Stanford, or anyone else.
Confidence Scoring
Each detection gets a confidence score from 0.0 to 1.0. This helps with human review decisions. Highly specific formats score high. For example, the German IBAN DE89 3704 0044 0532 0130 00 scores 0.85 or higher. Generic digit patterns score lower, around 0.3 to 0.5. These rely on context words to confirm a match. Compliance teams can review and override any detection before anonymization.
Context Word Analysis
Each recognizer has context words in its own language. For example, 'Personalausweis' works for German IDs, and 'kitambulisho' works for Kenyan IDs. When a context word appears near a match, the confidence score goes up.
Supported Entity Types
Full coverage of personal information types, organized into clear categories.
Personal Identifiers
- Person Names
- Email Addresses
- Phone Numbers
- Date of Birth
- Age
- Gender
- Nationality
Financial Information
- Credit Card Numbers
- IBAN
- BIC/SWIFT
- Bank Account Numbers
- Tax IDs
- VAT Numbers
Government IDs
- Social Security Numbers (SSN)
- National ID Numbers
- Passport Numbers
- Drivers License
- Health Insurance IDs
Location Data
- Street Addresses
- Cities
- ZIP/Postal Codes
- Countries
- GPS Coordinates
Digital Identifiers
- IP Addresses (v4/v6)
- MAC Addresses
- URLs
- Domain Names
- User IDs
Organization Data
- Company Names
- Organization IDs
- Registration Numbers
- Department Names
Temporal Data
- Dates
- Times
- Date Ranges
- Timestamps
International Formats
- German ID (Personalausweis)
- UK National Insurance
- Spanish DNI/NIE
- Italian Codice Fiscale
- And 70+ more country-specific formats
Custom Entity Support
Need to detect custom patterns? Create your own entity types with regex patterns. Or use our AI pattern generator instead.
Manual Pattern Creation
Define your own regex patterns for private identifiers. Examples include employee IDs, project codes, and reference numbers.
AI Pattern Generator
Describe what you want to detect in plain language. Our AI writes the regex patterns for you.
Is This Right For You?
Best For
- ✦Teams needing 317 entity types across personal, financial, government, and organizational data
- ✦Multilingual PII detection (48 languages) for global compliance and audit requirements
- ✦GDPR, HIPAA, or air-gapped environments requiring local processing and regex-first detection
- ✦Auditable detection workflows using hybrid regex + NLP + ML with confidence scoring
- ✦Structured and unstructured data (CSV, JSON, plain text, email) with custom entity patterns
Not For
- ✦Real-time processing with <10ms latency requirements (NLP models add ~50ms overhead)
- ✦Image-only PII detection workflows (use the dedicated image-redaction feature instead)
- ✦Simple English-only use cases where basic regex or keyword matching is sufficient