What is Microsoft Presidio?

An open-source SDK for fast PII detection and anonymization in text and images — it identifies entities like names, credit cards, SSNs, phone numbers, and crypto wallets, then replaces or transforms them.

Presidio (Latin praesidium, "protection/garrison") aims to democratize de-identification: give organizations customizable, transparent tools for both fully- and semi-automated PII workflows.

How it works in practice — feed it text like "Hi, my name is David Johnson from Liverpool, my card is 4095-2609-..." and it outputs "Hi, my name is <PERSON> from <LOCATION>, my card is <CREDIT_CARD>...". It detects PERSON, LOCATION, CREDIT_CARD, CRYPTO, EMAIL, IP_ADDRESS, IBAN, US_PASSPORT, and more, with a configurable acceptance threshold and pluggable NER models.

Try it: docs at microsoft.github.io/presidio and a live demo on Hugging Face Spaces (presidio/presidio_demo).