MCP PII Conceal Proxy

Example Use Cases

• 1

  Protect sensitive PII data before sending it to external AI providers.

• 2

  Consistent and realistic pseudo-anonymization to preserve data relationships.

• 3

  Flexible detection modes balancing speed and accuracy in PII detection.

Description

MCP Conceal acts as a proxy server that detects and pseudo-anonymizes PII in data streams to external AI services like Claude, ChatGPT, or Gemini. It preserves semantic meaning and data relationships by replacing real PII with consistent, realistic fake data, enabling safe AI analysis without exposing sensitive information. The server supports multiple detection modes including regex and LLM-based detection, configurable mappings, and integration with Ollama for advanced PII detection.