Redact MCP

Redact MCP is an MCP server—fully compatible with Claude Code and other MCP-enabled clients—that automatically detects and obfuscates sensitive data such as IPs, emails, API keys, and personal/organizational information. Using both regex patterns and AI-powered NER (Named Entity Recognition), it ensures client data is never exposed to the language model or AI assistant during workflows like code analysis and penetration testing. Its key feature is a secure HTTP proxy for pentesting or code review, allowing sensitive requests and responses to be sanitized on the fly, and supports full bidirectional mapping and audit trails for compliance. Redact MCP is ideal for penetration testers, security and compliance professionals, and anyone integrating AI into workflows that handle PII or sensitive data.