PII Scrubber for AI Tools

The tool detects: email addresses, phone numbers (US and international formats), Social Security Numbers, credit card numbers (Luhn-validated), dates, IP addresses, URLs, full names (capitalized pairs), street addresses, and ZIP/postal codes. You can also add custom words or phrases — company names, project names, or any specific term you want removed.

Instead of removing PII entirely, this mode replaces each detected item with realistic-looking but completely fake data: real emails become fake@example.com, phone numbers become (555) 000-0000 style numbers, names become generic placeholders like "Alex Johnson", and SSNs become 000-00-0000. This preserves the structure and readability of your text while removing all real personal data — useful when you need an AI to understand the context of a document without exposing actual personal details.

No. All detection, replacement, and tokenization runs entirely in JavaScript in your browser tab. Your text never leaves your device. There are no API calls, no analytics on your content, and no logging. The tool works fully offline once the page is loaded.

Public AI models like ChatGPT (OpenAI) and Claude (Anthropic) may use submitted text for model training by default depending on your plan settings. Even if they do not, submitting customer data, employee records, legal documents, or medical notes to third-party AI services may violate GDPR, HIPAA, or your organisation's data handling policies. Scrubbing PII before pasting lets you use AI tools productively without compliance risk.

Tokenization replaces each detected PII item with a reversible placeholder like [EMAIL_1], [PHONE_2], [NAME_1]. The detection map shows which token maps to which original value, so you can ask the AI a question using tokenized text, get a response with the same tokens in it, and then swap the real values back in yourself. This is the most privacy-preserving mode for workflows that require the AI to process structured data.