#14 · Primary category： Privacy-Preserving & Federated Data Science
rizzo-pii
Local-first privacy guard: anonymize your documents before sharing with LLMs.
Project last updated：08/09/26
GitHub Stars
830
Forks
57
Contributors
14
License
MIT
Why we included this project
Anyone working with Italian legal or financial documents faces a real dilemma: the most capable hosted models are exactly the ones you can't safely send client data to. rizzo-pii runs a small token-classification model on your CPU to tag 22 categories of personal data, including the codice fiscale, partita IVA, and cadastral references that English-first detectors tend to miss. Each value is swapped for a stable placeholder before anything leaves the machine, and the mapping back to the real data stays in a local dictionary, so a later pass restores the true names and numbers once the cloud model has answered. That reversibility is the point: the document stays usable for real work, where plain redaction would just destroy the information. Law firms, accountants, and notaries get GDPR data minimization without giving up the reasoning quality of a large hosted model.
Articles for this project
No articles for this project yet.
To suggest a topic or contribute an article, contact us.
Related projects in this category
watermarks-remover
A privacy-first app that strips AI watermarks from content you own.
PySyft
Perform data science on data that remains in someone else's server
presidio
An open-source framework for detecting, redacting, masking, and anonymizing sensitive data (PII) across text, images, and structured data. Supports NLP, pattern matching, and customizable pipelines.
flower
Flower: A Friendly Federated AI Framework
ai.robots.txt
A list of AI agents and robots to block.