chiefautism

chiefautism / privacy-parser

Public

Reverse of OpenAI Privacy Filter: same 1.5B model, returns PII as structured spans instead of masking.

340
30
89% credibility
Found Apr 27, 2026 at 340 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A Python tool that extracts structured personally identifiable information (PII) like names, emails, phones, and secrets from text using regex patterns, optionally enhanced by an OpenAI-compatible AI model.

How It Works

1
🔍 Discover Privacy Parser

You hear about a helpful tool that spots personal details like names, emails, and phone numbers hidden in text, perfect for checking your own data or reviewing leaks.

2
📥 Get the Tool Ready

You download the tool and set it up on your computer so it's ready to use right away.

3
Pick Your Scanning Style
⚡
Quick Scan

Uses simple patterns for super-fast results on basic personal info.

🧠
Smart Scan

Adds clever thinking to catch tricky details, grabbing a helper file the first time.

4
📝 Feed It Your Text

Paste or type in the text you want to check, like an email or log file.

5
✨ See the Magic Happen

The tool quickly highlights every bit of personal info it finds, labeling names, addresses, secrets, and more.

6
📋 Review Your List

Get a neat list of all discovered details with their exact spots in the text.

✅ Personal Info Secured

You now have a clear view of sensitive details to protect your data or analyze leaks safely.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 340 to 340 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is privacy-parser?

Privacy-parser flips OpenAI's Privacy Filter on its head: it uses the same 1.5B model to pull out PII like emails, phones, and secrets as structured spans—label, position, and exact text—instead of just masking them. Built in Python, it offers a simple CLI for quick scans (`python -m pii_parser "your text"`) and a clean API for batch processing logs or dumps. Perfect for auditing data before it leaks or extracting from breaches, without changing your OpenAI reverse proxy setup.

Why is it gaining traction?

340 stars show devs digging its drop-in compatibility with OpenAI's filter schema, letting you swap masking for extraction seamlessly. The hybrid backend nails 0.929 F1 score by blending the model with regex fallbacks, beating pure model (0.733) on speed and accuracy—600ms on CPU. CLI spits JSON or spans, and first-run auto-downloads the 3GB checkpoint.

Who should use this?

Security engineers sifting S3 leaks or logs for PII cleanup. Pentesters parsing stolen inboxes with precise spans. Devs routing Azure OpenAI via reverse proxy who need structured privacy checks before prompts hit the API.

Verdict

Grab it for production PII extraction—Apache-2.0, solid benchmarks, and CLI make it ready now, despite 0.1.0 version and 0.8999999761581421% credibility score. With 340 stars and passing tests, it's mature enough for real workflows; watch for more backends.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.