Entity Types

Overview

@secured-ai/core defines 44 entity types as a TypeScript string union. Every detected entity is assigned one of these types, and you use them when configuring replacement pools, writing custom patterns, and filtering results.

import type { EntityType } from '@secured-ai/core'

The types are split into two groups:

Sensitive types (29) — treated as PII by obfuscate() and included in sensitiveEntities
Informational types (15) — detected by detect() but not automatically obfuscated

Sensitive types

These 29 types are classified as sensitive. obfuscate() will replace any entity of these types with a fake value.

Type	Description	Example	Detecting engine(s)
`PERSON`	Full name of a natural person	`Jane Doe`	NLP, ML
`EMAIL`	Email address	`jane@example.com`	Regex
`PHONE`	Phone number (various formats)	`555-867-5309`	Regex
`SSN`	US Social Security Number	`123-45-6789`	Regex
`CREDIT_CARD`	Credit or debit card number	`4111 1111 1111 1111`	Regex
`CVV`	Card verification value	`427`	Regex
`ORG`	Organisation or company name	`Acme Corp`	NLP, ML
`GPE`	Geopolitical entity (city, country, state)	`New York`	NLP, ML
`LOC`	Non-GPE location (building, landmark)	`Golden Gate Bridge`	NLP, ML
`ADDRESS`	Street address	`123 Main St, Springfield`	Regex
`ZIP_CODE`	Postal code	`90210`	Regex
`DATE`	Calendar date	`January 15, 1985`	Regex, NLP
`TAX_ID`	Tax identification number	`12-3456789`	Regex
`ACCOUNT_NUMBER`	Bank account number	`000123456789`	Regex
`INVOICE_NUMBER`	Invoice or billing reference	`INV-2024-0042`	Regex
`REFERENCE_ID`	Generic reference identifier	`REF-US-1234`	Regex, Custom
`MEDICAL_ID`	Medical record number	`MRN-00123456`	Regex
`INSURANCE_ID`	Insurance policy number	`POL-4521-ABC`	Regex
`CASE_NUMBER`	Legal or support case number	`CASE-2024-00789`	Regex
`LEGAL_CITATION`	Legal case citation	`Brown v. Board of Education`	Regex
`DRIVER_LICENSE`	Driver's license number	`DL-1234567`	Regex
`PASSPORT`	Passport number	`A12345678`	Regex
`IP_ADDRESS`	IPv4 or IPv6 address	`192.168.1.1`	Regex
`MAC_ADDRESS`	Network hardware address	`00:1A:2B:3C:4D:5E`	Regex
`URL`	Web URL	`https://example.com/user/123`	Regex
`ROUTING_NUMBER`	US bank routing number	`021000021`	Regex
`IBAN`	International Bank Account Number	`GB29NWBK60161331926819`	Regex
`EIN`	Employer Identification Number	`12-3456789`	Regex

Informational types

These 15 types are detected by detect() and appear in result.entities, but are not in sensitiveEntities and are not replaced by obfuscate() by default.

Type	Description	Example	Detecting engine(s)
`CREDIT_CARD_EXPIRY`	Card expiry date	`12/28`	Regex
`TIME`	Time of day	`3:45 PM`	Regex
`NORP`	Nationality, religion, or political group	`American`, `Buddhist`	NLP
`FAC`	Facility name	`JFK Airport`	NLP
`PRODUCT`	Product or service name	`iPhone 15 Pro`	NLP
`EVENT`	Named event	`World War II`, `Super Bowl`	NLP
`WORK_OF_ART`	Title of a creative work	`The Great Gatsby`	NLP
`LAW`	Named law or regulation	`GDPR`, `HIPAA`	NLP
`LANGUAGE`	Language name	`Spanish`, `Mandarin`	NLP
`QUANTITY`	Measurement or quantity	`500 metres`, `3 kilograms`	NLP
`CARDINAL`	Cardinal number	`forty-two`	NLP
`ORDINAL`	Ordinal number	`first`, `third`	NLP
`PERCENT`	Percentage	`87%`	NLP
`MISC`	Miscellaneous named entity	Various	ML
`CUSTOM`	User-defined entity type	Domain-specific	Custom

Using entity types in code

Filtering detection results

import type { EntityType } from '@secured-ai/core'

const result = await client.detect(text)

const namesOnly = result.entities.filter(
  (e) => e.type === 'PERSON'
)

const financialPII = result.entities.filter(
  (e): e is typeof e & { type: EntityType } =>
    ['CREDIT_CARD', 'IBAN', 'ROUTING_NUMBER', 'ACCOUNT_NUMBER'].includes(e.type)
)

Custom replacement pools by type

const client = new PrivacyClient({
  baseUrl: 'https://dev-api.securedai.com',
  sdkAccessToken: import.meta.env.VITE_SECURED_SDK_ACCESS_TOKEN,
  replacementPools: {
    PERSON: ['Alex Taylor', 'Jordan Lee', 'Morgan Smith'],
    EMAIL: ['user@redacted.example'],
    PHONE: ['800-555-0100'],
    SSN: ['000-00-0000'],
  },
})

Custom patterns with specific types

client.addPattern({
  name: 'employee-id',
  pattern: /\bEMP-\d{6}\b/,
  entityType: 'REFERENCE_ID',  // classify as a sensitive REFERENCE_ID
  confidence: 0.97,
})

Notes on `CUSTOM`

The CUSTOM type is a catch-all for user-defined patterns. Note that CUSTOM is not in the sensitive types list — entities matched with entityType: 'CUSTOM' will appear in detect() results but will not be replaced by obfuscate(). Use a built-in sensitive type (like REFERENCE_ID or CASE_NUMBER) if you want automatic obfuscation.

Entity Types

On this page