NLP Overview

When to use NLP

The built-in detectors use regex patterns and key-name heuristics. They work well for structured values like emails, phone numbers, and government IDs.

For freeform text containing person names and place names that don't follow a pattern, use @pii-mask/nlp. It wraps the compromise NLP library to detect entities.

Accuracy caveats

NLP detection is probabilistic, not deterministic. It may:

• Miss uncommon names
• Falsely detect common words as names
• Perform differently across languages

Always test with representative data before deploying.

Installation

npm install @pii-mask/nlp compromise