Introduction

Worka PII is a deterministic PII detection and anonymization library designed to run on CPU-only environments without sacrificing clarity or auditability. It was built to meet the needs of AI applications where redaction must be reliable, repeatable, and easy to reason about.

Unlike probabilistic or opaque pipelines, Worka PII emphasizes explicit policies, predictable offsets, and controlled degradation. If a language lacks NER or lemma support, the pipeline still operates with the available capabilities instead of failing or silently changing behavior.

Use Worka PII when you need:

• Consistent detection boundaries and stable byte offsets.
• A pipeline you can tune with explicit policies and recognizers.
• Clear audit logs that show what changed and why.
• A safe default for AI prompts, logs, and data exports.