data-privacy-stack/presidio ? reverse-engineered prompt

Build me a Python privacy toolkit that can find and hide sensitive information in text, images, and structured data.

I want it to detect things like names, phone numbers, email addresses, credit cards, SSNs, and similar PII, then let me redact, mask, replace, or anonymize them. It should work with both simple pattern matching and smarter language based detection, and I should be able to customize the rules for my own use case.

Please include a clean way to use it from Python, plus a command line interface for quick jobs. It would be great if it also supports redacting text inside images, including scanned or medical style images, and has a path for structured data like tables or records. Keep it easy to run locally, and if you need to check current docs or package details online, go ahead and do that. I’d like the result to feel like a real open source library that is practical, configurable, and safe to use in production style workflows.