Detect and mask PII for GDPR-safe AI document analysis with Anthropic and PostgreSQL

Overview

This workflow enables GDPR-compliant document processing by detecting, masking, and securely handling personally identifiable information (PII) before AI analysis.

It ensures that sensitive data is never exposed to AI systems by replacing it with tokens, while still allowing controlled re-injection of original values when permitted. The workflow also maintains full audit logs for compliance and traceability.

---

How It Works

1. Document Upload & Configuration Receives documents via webhook and initializes configuration such as document ID, thresholds, and database tables.

2. Text Extraction Extracts raw text from uploaded documents for processing.

3. Multi-Detector PII Detection Detects emails, phone numbers, ID numbers, and addresses using regex and AI-based detection.