Repolode
Submit
data-privacy-stack

data-privacy-stack/presidio

Open-source SDK for detecting and anonymizing PII in text, images and structured data, using NLP, regex and rules.

Open-source alternative to:Amazon MacieGoogle Cloud Data Loss Prevention

View on GitHubWebsite
Stars
11,388
Forks
1,357
Open issues
197
Last commit
20 hours ago

What is it?

Presidio is a Python framework for finding and de-identifying personally identifiable information (PII). It identifies entities such as credit card numbers, names, locations, social security numbers, bitcoin wallets and US phone numbers, then anonymizes them through redaction, masking or other transformations.

Detection combines named entity recognition, regular expressions, rule-based logic and checksum validation, with context words considered across multiple languages. Custom recognizers can be added, and external PII detection models can be connected. The project is split into modules: an analyzer, an anonymizer, an image redactor (including DICOM medical images) and a structured-data component.

Presidio can be used from Python or PySpark, run in Docker, or deployed on Kubernetes. The project notes that automated detection cannot guarantee every sensitive item is found, so it should be combined with other safeguards.

Key features

  • 01Detects predefined or custom PII entities using NER, regex, rules and checksums.
  • 02Supports context-aware recognition across multiple languages.
  • 03Redacts PII in text and in images, including DICOM medical images.
  • 04Can connect to external PII detection models.
  • 05Runs from Python, PySpark, Docker or Kubernetes.

Use cases

  • Scrubbing personal data from logs, documents or datasets before sharing or analysis.
  • De-identifying medical images or records in healthcare pipelines.
  • Adding PII guardrails to LLM inputs and outputs.
  • Automating redaction of sensitive fields in large data processing jobs with PySpark.

Topics

#pii-detection#data-privacy#anonymization#data-masking#nlp#named-entity-recognition#python#de-identification#data-anonymization#data-obfuscation

FAQ

What is presidio?
Open-source SDK for detecting and anonymizing PII in text, images and structured data, using NLP, regex and rules. Presidio is a Python framework for finding and de-identifying personally identifiable information (PII). It identifies entities such as credit card numbers, names, locations, social security numbers, bitcoin wallets and US phone numbers, then anonymizes them through redaction, masking or other transformations.
Is presidio free to use? What license does it have?
presidio is open source under the MIT license. Check the license terms before using it in commercial products.
Is presidio still maintained?
Actively maintained. The last commit to presidio was on Oct 18, 2026.
What can I use presidio for?
Scrubbing personal data from logs, documents or datasets before sharing or analysis. De-identifying medical images or records in healthcare pipelines. Adding PII guardrails to LLM inputs and outputs. Automating redaction of sensitive fields in large data processing jobs with PySpark.
Where can I get presidio?
The source code and installation instructions are on GitHub: https://github.com/data-privacy-stack/presidio

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects