Portuguese is spoken by over 260 million people across four continents, with Brazil representing the largest Portuguese-speaking population by far, followed by Portugal, Mozambique, Angola, and other lusophone nations. While these countries share a common language, they differ significantly in vocabulary, spelling conventions, naming practices, identifier formats, and address structures. Effective Portuguese PII detection must account for these regional variations while maintaining consistent protection across all variants.
Our Portuguese language processing handles both major variants—Brazilian Portuguese (pt-BR) and European Portuguese (pt-PT)—with dedicated models for each. We recognize regional naming conventions, validate country-specific identifiers with appropriate algorithms, and parse address formats used in each country. This enables comprehensive PII protection for Portuguese-language documents regardless of their origin.
Regional Variants
Key differences between Brazilian and European Portuguese:
Vocabulary Differences:
PII-related terms:
Identity document: carteira de identidade (BR) vs bilhete de identidade (PT)
Cell phone: celular (BR) vs telemóvel (PT)
Address: endereço (both, but formatting differs)
Personal data: dados pessoais (both)
Passport: passaporte (both)
Driver's license: carteira de motorista (BR) vs carta de condução (PT)
// Detection handles both vocabularies
"Telemóvel: 912 345 678" → [PHONE] (Portugal)
"Celular: (11) 99999-8888" → [PHONE] (Brazil)
Spelling Differences:
// Post-1990 agreement variations
ação (BR) vs acção (PT, older)
anônimo (BR) vs anónimo (PT)
fato (BR) vs facto (PT)
// Names may follow either convention
Both spellings recognized for matching
Portuguese Name Detection
Portuguese naming conventions across regions:
Brazilian Names:
Structure: Given name(s) + Maternal surname + Paternal surname
Example: Maria Eduarda Silva Santos
Components:
- Given names: Often religious or Italian/Portuguese origin
- Multiple given names common
- Surnames: Usually maternal then paternal
- Sometimes only one surname
Common patterns:
João Pedro Oliveira Costa
Ana Carolina Ferreira Lima
José Carlos da Silva
Maria das Graças Santos
Portuguese (European) Names:
Structure: Similar but with different common names
Example: António Manuel Rodrigues Sousa
Components:
- Given names: Traditional Portuguese names
- Surnames: Often preceded by "de", "da", "dos"
- Nobility particles preserved
Common patterns:
António José Almeida
Maria João Carvalho
Pedro Miguel Teixeira da Costa
Ana Sofia Santos e Silva
Name Detection Approach:
// Comprehensive name databases
Brazilian first names: João, José, Maria, Ana, Pedro, Paulo...
Portuguese first names: António, Manuel, Joaquim, Rui...
Shared surnames: Silva, Santos, Oliveira, Souza/Sousa...
// Context indicators
Nome: [name]
Nome completo: [full name]
Cliente: [name]
Sr./Sra./Dr./Dra. [name]
Country-Specific Identifiers
Brazilian Identifiers:
CPF (Cadastro de Pessoas Físicas):
Format: XXX.XXX.XXX-XX (11 digits)
Example: 123.456.789-09
Validation: Modulo 11 checksum
CNPJ (Cadastro Nacional da Pessoa Jurídica):
Format: XX.XXX.XXX/XXXX-XX (14 digits)
Example: 11.222.333/0001-81
Validation: Modulo 11 checksum
RG (Registro Geral):
Format: Varies by state
Example: 12.345.678-9 (SSP-SP)
Portuguese Identifiers:
NIF (Número de Identificação Fiscal):
Format: XXXXXXXXX (9 digits)
Example: 123456789
First digit indicates entity type:
- 1, 2, 3: Individual
- 5: Corporate
- 6: Public administration
CC (Cartão de Cidadão):
Format: XXXXXXXXX XX X (12 characters)
Example: 12345678 9 ZZ4
Components: Base number + Check digits + Version
NISS (Número de Identificação de Segurança Social):
Format: XXXXXXXXXXX (11 digits)
Example: 12345678901
Other Lusophone Countries:
Angola:
- NIF: Tax identification
- BI: Bilhete de Identidade
Mozambique:
- NUIT: Tax identification
- BI: Identity document
Cape Verde:
- NIF: Tax number
- BI: Identity card
Address Formats
Brazilian Addresses:
Format:
[Street type] [Name], [Number] - [Complement]
[Neighborhood]
[City] - [State]
CEP: [Postal code]
Example:
Rua das Flores, 123 - Apto 45
Jardim Paulista
São Paulo - SP
CEP: 01310-100
CEP format: XXXXX-XXX (8 digits)
Regions indicated by first digits
Portuguese Addresses:
Format:
[Street type] [Name], [Number], [Floor]
[Postal code] [City]
Example:
Rua Augusta, 25, 3º Esq.
1100-053 Lisboa
Código Postal format: XXXX-XXX (7 digits)
First 4: Region, Last 3: Specific location
Street types: Rua, Avenida, Praça, Largo, Travessa
Phone Number Formats
Brazilian Phone Numbers:
Mobile: (XX) 9XXXX-XXXX (11 digits with area code)
Landline: (XX) XXXX-XXXX (10 digits)
International: +55 XX XXXXX-XXXX
Examples:
(11) 99999-8888 // São Paulo mobile
(21) 2222-3333 // Rio de Janeiro landline
+55 11 99999-8888 // International format
Area codes: 11-99 across states
Portuguese Phone Numbers:
Mobile: 9XX XXX XXX (9 digits, starts with 9)
Landline: 2XX XXX XXX (9 digits, starts with 2)
International: +351 XXX XXX XXX
Examples:
912 345 678 // Mobile
213 456 789 // Lisbon landline
+351 912 345 678 // International
Mobile prefixes: 91, 92, 93, 96
Landline prefixes: 21 (Lisbon), 22 (Porto), etc.
Financial Identifiers
Brazilian Banking:
Bank account format:
Banco: [3-digit bank code]
Agência: [4 digits]-[check digit]
Conta: [variable digits]-[check digit]
Example:
Banco: 001 (Banco do Brasil)
Agência: 1234-5
Conta: 123456-7
PIX keys: CPF, phone, email, or random key
Portuguese Banking:
NIB (Número de Identificação Bancária):
Format: XXXXXX.XXXX.XXXXXXXXXXX.XX (21 digits)
Example: 0035.0123.00001234567.89
IBAN format:
PTXX XXXX XXXX XXXX XXXX XXXX X (25 characters)
Example: PT50 0035 0123 0000 1234 5678 9
Regulatory Compliance
Brazil - LGPD:
- Lei Geral de Proteção de Dados (effective 2020)
- Similar structure to GDPR
- Enforced by ANPD (National Data Protection Authority)
- Significant fines for non-compliance
Portugal - GDPR + National Law:
- GDPR directly applicable (EU member)
- Lei 58/2019 implementing GDPR
- CNPD (Comissão Nacional de Proteção de Dados)
- Sector-specific regulations
Language Detection and Processing
// Automatic variant detection
Input: "O cliente João Silva forneceu seu telemóvel"
Detected: pt-PT (telemóvel indicates European Portuguese)
Input: "O cliente João Silva forneceu seu celular"
Detected: pt-BR (celular indicates Brazilian Portuguese)
// Processing adapts to variant
- Identifier formats appropriate to region
- Address parsing matching regional format
- Name patterns for detected region
// Can also specify explicitly
{
"language": "pt",
"region": "BR" // or "PT"
}
API Configuration
POST /v1/redact
{
"text": "Cliente: Maria Silva, NIF: 123456789, Telemóvel: 912 345 678",
"language": "pt",
"auto_detect_region": true,
"redaction_types": ["name", "tax_id", "phone", "address"]
}
Response:
{
"redacted_text": "Cliente: [NAME], NIF: [TAX_ID], Telemóvel: [PHONE]",
"detected_region": "PT",
"detections": [
{
"type": "name",
"value": "Maria Silva",
"confidence": 0.94
},
{
"type": "tax_id",
"value": "123456789",
"format": "portugal_nif",
"confidence": 0.97
},
{
"type": "phone",
"value": "912 345 678",
"format": "portugal_mobile",
"confidence": 0.98
}
]
}
