presidio

Why we included this project

Presidio is a solid pick when you need to strip personally identifiable information out of data before it goes anywhere it shouldn't. It finds PII in free text using a mix of named-entity recognition, regex, rule-based checks, and checksum validation, so it catches credit card numbers, names, locations, and social security numbers without relying on any single method. The same detection core also handles images, including DICOM medical files, and structured data, so one framework covers text, documents, and tabular exports. Recognizers are pluggable and the pipeline is customizable, which means you can add your own entity types or plug in external detection models instead of living with the defaults. It runs from Python or PySpark and deploys through Docker or Kubernetes, so it scales from a small internal guardrail to a service-wide de-identification layer.