Presidio is an open-source data protection and de-identification SDK. It detects personal data (PII) in text, images and structured data and can redact or anonymise it. The Analyzer combines named entity recognition, regular expressions, rules and checksums; the Anonymizer applies the chosen operation; the Image Redactor masks text in images, including DICOM medical images. You use Presidio from Python or PySpark, through Docker or on Kubernetes. According to the project, automated detection cannot guarantee that all sensitive data is found.
Presidio is MIT-licensed and run by the user: there is no vendor-hosted service, so data stays in your own environment. The project was started by Microsoft and now sits under the Data Privacy Stack organisation, which describes the change as a transition to a community-owned project. For a European organisation this means the code can be inspected and self-hosted, while the project's governance is still in transition.