# PII & Secret Redactor [Health: Active]

**Category:** 🔒 Security  
**Repository:** https://github.com/tylerscomic-lab/pii-redactor-mcp  
**GitHub Stars:** 0  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/pii-secret-redactor

## Description
Detect and redact PII and secrets before text reaches an LLM, with reversible placeholders.

## Tools
Capabilities this server exposes over MCP:

- **detect_pii** — findings with type, position and a masked preview. Never echoes full values.
- **redact_text** — placeholder (stable tokens like `<EMAIL_1>`, same value gives the same token), mask, salted hash, or remove.
- **restore_text** — swap placeholders back to the originals in the model's response, using the mapping you keep.
- **list_detectors** — exactly what is and is not covered.

## Claude Desktop Quick Installation
No confirmed install details yet, so no config is published. Follow the setup steps in the repository above.

## Documentation

## What PII & Secret Redactor MCP server does

The PII & Secret Redactor MCP server places a detection and transformation step in front of an LLM workflow. It can inspect text for contact details, identity data, network addresses, financial identifiers, and several classes of credentials. The server is aimed at reducing exposure when prompts, logs, or support tickets may contain customer data or secrets.

Detection includes email addresses, US and international phone numbers, labelled dates of birth, US street addresses, IPv4 and IPv6 addresses, and selected payment and banking identifiers. Credit cards are checked with the Luhn algorithm and their brand is identified. IBANs use mod-97 validation, US routing numbers use an ABA checksum, and SSNs are checked against invalid ranges. The detector also covers API keys from Anthropic, OpenAI, AWS, GitHub, Stripe, Slack, and Google, along with JWTs, private keys, and credential-bearing database URLs.

## How it works

The `detect_pii` tool reports each finding’s type and position with a masked preview. It does not echo the complete matched value. `redact_text` applies one of four transformations: stable placeholders such as `<EMAIL_1>`, masking, salted hashing, or removal. Stable placeholders allow repeated occurrences of the same value to use the same token within the redaction operation.

For reversible workflows, retain the mapping produced by redaction and pass it to `restore_text` after the model returns an answer. That tool replaces placeholders with their original values. `list_detectors` provides the authoritative inventory of supported and unsupported detector types, which is useful for checking coverage before relying on the server in a pipeline.

## Setup and configuration

The server is available as a hosted MCP endpoint through MCPize. The remote streamable HTTP URL is `https://pii-redactor-mcp.mcpize.run/mcp`, and access uses an API key from MCPize. The hosted service has a free tier limited to 10 calls per day.

For a local run, install the package with `npm install`, then start it with `node server.js`. The process listens on port 8080 and exposes MCP at `/mcp`. No environment variables are specified in the provided setup instructions.

## Tools and capabilities

- `detect_pii`: finds supported PII and secret patterns, returning types, positions, and masked previews.
- `redact_text`: transforms matched values using placeholders, masks, salted hashes, or removal.
- `restore_text`: restores original values from placeholders using a mapping retained by the caller.
- `list_detectors`: shows what the detector covers and what it does not.

The PII & Secret Redactor MCP server processes input in memory and does not store or log it, according to the project documentation. It is listed in the Security category and is MIT licensed.

## Limitations and notes

Detection is based on patterns and checksums. Personal names and free-form addresses in formats outside the supported rules are not detected. Consequently, the PII & Secret Redactor MCP server can reduce accidental exposure but does not establish HIPAA, GDPR, or PCI compliance. Review `list_detectors` and test representative inputs before using it as part of a production data-handling process.

_Full upstream README: https://allmcps.com/mcp/pii-secret-redactor/readme_