Anonymize
Deprecated
Use phi-deidentify node instead. It provides NLP-based detection and FHIR-aware processing.
Category: Security
Node id: anonymize
Maturity: beta
Risk tier: B
Standards: HIPAA, SOC2
Removes or masks PII/PHI from text and structured data using redaction, masking, hashing, or fake data generation
What it does
- Four anonymization methods: redact, mask, hash, fake
- Free-text and structured data anonymization
- Hash-based anonymization with configurable salt
- Replacement tracking with count reporting
- Additional field targeting via config
Use cases
- Anonymize clinical notes before research use
- Hash patient identifiers for pseudonymized datasets
- Redact PHI from structured FHIR resources before export
Runtime
- Performance tier:
fast
Input ports
| Name | Type | Required | Description |
|---|---|---|---|
text | text | No | Free-text to anonymize |
structured_data | json | No | Structured data map to anonymize |
Output ports
| Name | Type | Required | Description |
|---|---|---|---|
result | json | No | Anonymization output containing anonymized_text, anonymized_data, replacements, and replacement_count |
success | boolean | No | Whether anonymization succeeded |
Settings
Basic
| Setting | Name | Type | Required | Default | Allowed values | Description |
|---|---|---|---|---|---|---|
| Method | method | enum | Yes | redact | redact, mask, hash, fake | Anonymization method to apply |
Auth
| Setting | Name | Type | Required | Default | Allowed values | Description |
|---|---|---|---|---|---|---|
| Hash Salt | hash_salt | string | No | tietai-anonymize | Salt value for hash-based anonymization |
Advanced
| Setting | Name | Type | Required | Default | Allowed values | Description |
|---|---|---|---|---|---|---|
| Extra Fields | fields | string | No | Comma-separated additional field names to anonymize in structured data |
Errors
| Code | Description |
|---|---|
INVALID_METHOD | The specified anonymization method is not recognized |
MISSING_INPUT | Neither text nor structured_data input was provided |
Runtime notes
- Replacement is regex-driven; text the patterns do not match passes through unchanged.
- result.anonymized_text is only present when text input is provided; result.anonymized_data only when structured_data is provided
- The replacements array records every substitution made — use it for audit trails or reversibility with hash method
- Hash method with the same salt produces deterministic output, enabling pseudonymized joins across datasets
- Structured data anonymization recognizes common PHI field names (name, ssn, dob, etc.) automatically
See Build any workflow for how to put nodes together and Ports, types and data flow for how port types connect.
Something unclear? Send documentation feedback by email.