piigrep

A command that walks a directory like grep -r and reports the personal information it finds in files.

Each find is reported as path:line:kind:value. The default is a fast pattern search, and on macOS 27 or later the on-device Apple Foundation Model can also find names, addresses, and birthdays. The model runs on the Mac, and nothing is sent anywhere.

Features

• Patterns cover email addresses, Japanese phone numbers, postal codes, My Numbers, and credit card numbers. My Numbers must pass their check digit and card numbers the Luhn check
• Full-width digits are understood, and values are reported as written
• Reads UTF-8, Shift_JIS, and UTF-16 with a BOM. Binary files are skipped
• Hidden files and files listed in .gitignore are searched by default, since .env files and dumps are where personal information tends to sit (--gitignore respects .gitignore)
• --format json / tsv for machine-readable output

Requirements & notes

• The patterns are for Japanese formats. Other countries' ID numbers are not looked for.
• --fm requires macOS 27 or later with Apple Intelligence turned on.
• Binaries are published for macOS on Apple Silicon (signed and notarized) and x86_64 Linux (glibc).
• Open source under the MIT License.

Links

• Product page: https://www.cyberneura.com/app/piigrep
• GitHub: https://github.com/cyberneura/piigrep