Skip to main content
Automatically detect and redact personally identifiable information in agent requests and responses. This policy covers the four most common PII patterns and redacts them before they reach the LLM or the end user.

Policy JSON

What It Does

  • Scans both input and output for email addresses, phone numbers, Social Security Numbers, and credit card numbers
  • Redacts matches by replacing them with [REDACTED_EMAIL], [REDACTED_PHONE], etc.
  • Redacts in audit logs so PII never persists in your logging infrastructure
  • Uses deterministic regex-based detection — no LLM calls in the governance path

Example

User input:
After TealGuard processing:
Decision object:

Install

Customize

Setting action to DENY will block the entire request if any PII is detected. Use REDACT for most use cases to allow the request to proceed with PII removed.