# PII Redaction Deterministic Scrubber MCP for AI Agents MCP > PII Redaction Deterministic Scrubber identifies and masks sensitive data like SSNs, emails, phone numbers, credit cards, and IBANs. It uses strict regular expressions and checksum validators to ensure high-precision scrubbing of unstructured text without the guesswork of probabilistic models. ## Overview - **Category:** security - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_xqFNBPuR3qBDw1MH8ymNDz5FH6VBr9xTyyZ6CACx/mcp - **Tags:** pii, redaction, security, regex, privacy, compliance, sanitization ## Description Cleaning sensitive data shouldn't be a guessing game. When you are feeding logs, customer support transcripts, or unstructured documents into your AI client, you cannot afford to leave an email address or a credit card number exposed. This MCP acts as a high-precision filter that identifies and scrubs personally identifiable information before it ever reaches your agent. Instead of relying on models that might miss something, this tool uses exact patterns and mathematical checks like the Luhn algorithm to verify that financial identifiers are actually valid before redacting them. It does not just hide the data; it provides a structured breakdown of exactly what was found and where, so you can audit your privacy coverage without seeing the sensitive details themselves. You can connect this directly to your existing workflow through the Vinklan catalog, making it easy to add a layer of security to any automated process. It is about turning messy, risky text into clean, safe data that you can use for training or analysis without the compliance headache. ## Tools ### redact_pii Scans your input text to find and mask sensitive entities like SSNs or emails. It returns the cleaned text and a precise count of every match found. ## Prompt Examples **Prompt:** ``` Can you check this text for any sensitive info? 'Contact me at john.doe@example.com or call 555-0199.' ``` **Response:** ``` **Redaction Results:** **Redacted Text:** "Contact me at [REDACTED] or call [REDACTED]." **Detected Entities:** * email: 1 * phone: 1 ``` **Prompt:** ``` I have a list of credit card numbers in this log. Can you scrub them? ``` **Response:** ``` **Status:** Scrubbing complete. **Original Data Found:** 4532015112830363 **Action:** Masked to [REDACTED] **Verification:** Luhn checksum validated. ``` **Prompt:** ``` Scan this IBAN for any PII leaks: 'DE89370400440532013000' ``` **Response:** ``` **Result:** **Redacted Text:** [REDACTED] **Type:** IBAN **Match Found:** 1 ``` ## Capabilities ### Scrub sensitive identifiers Finds and masks SSNs, email addresses, and phone numbers in your text. ### Validate financial data Uses checksums to confirm the accuracy of credit card and IBAN numbers. ### Audit redaction coverage Provides exact character offsets so you can track every identified match. ### Quantify PII density Returns a precise count of each type of sensitive information detected. ### Sanitize unstructured text Cleans raw logs or transcripts to make them safe for downstream processing. ## Use Cases ### Cleaning Support Logs An engineer sends raw logs to an agent, and the MCP masks all customer emails and phone numbers automatically. ### Secure Dataset Preparation A researcher uses the tool to strip IBANs and SSNs from a large text corpus before training a new model. ### Automated Compliance Audits A compliance officer runs transcripts through the MCP to ensure no credit card numbers were leaked in chat logs. ### Sanitizing User Feedback A product manager uses the tool to scrub names and emails from user surveys before sharing them with the broader team. ## Benefits - Eliminate guesswork with deterministic patterns that do not rely on probability. - Verify financial accuracy using Luhn algorithm checks for credit cards. - Audit your privacy coverage by checking the exact character offsets returned. - Automate compliance workflows by sanitizing text before it reaches your agent. - Track PII density with detailed counts of every detected email or SSN. ## How It Works The bottom line is you get safe, sanitized text ready for any downstream use. 1. Connect the MCP to your preferred AI client via Vinkius. 2. Pass your unredacted text through the redaction process. 3. Receive a clean version of the text alongside a detailed report of what was masked. ## Frequently Asked Questions **How does PII Redaction Deterministic Scrubber handle SSNs?** It uses strict regular expression patterns to identify the specific format of a Social Security Number. This ensures that every instance is caught and masked accurately. **Can I use PII Redaction Deterministic Scrubber to clean logs for Claude?** Yes, you can connect this MCP to any compatible client like Claude or Cursor. It works perfectly for sanitizing logs before they are processed by your agent. **Does PII Redaction Deterministic Scrubber verify credit card numbers?** It does more than just find the pattern; it uses the Luhn algorithm to check if the number is a valid credit card. This prevents unnecessary redactions of non-sensitive number strings. **undefined** undefined **What kind of information does PII Redaction Deterministic Scrubber mask?** The tool specifically targets SSNs, email addresses, phone numbers, credit card numbers, and IBANs. It is designed for these high-risk identifiers. **Is the redaction process in PII Redaction Deterministic Scrubber reversible?** No, the masking is permanent. Once the text is processed, the sensitive information is replaced with a redacted placeholder, making it safe for use. **Can I audit what was removed using PII Redaction Deterministic Scrubber?** Yes, the tool returns character offsets and counts. This allows you to verify exactly which parts of your text were modified without ever seeing the original sensitive data. **Does this tool use AI or Machine Learning for detection?** No. It uses deterministic regex patterns and checksum validators like the Luhn algorithm to ensure 100% predictable results without ML overhead. **What types of PII can be redacted?** The `redact_pii` tool detects SSNs, emails, phone numbers, credit cards (via Luhn), and IBANs. **How can I use this in my AI agent?** You can connect via Vinkius Edge using your personal Connection Token. Once connected, you can call the `redact_pii` tool directly from Cursor, Claude Desktop, or any MCP-compatible client.