Features
Auto-detect over 30 PII types (names, emails, credit cards, API keys, passwords, JWTs, etc.)
Replace sensitive data with placeholders ([[PERSON_1]], [[EMAIL_1]], etc.)
Restore placeholders in AI responses where supported
Browser extension for ChatGPT, Claude, and Gemini (experimental)
HTTP API proxy for apps and SDKs (OpenAI/Anthropic compatible)
ENV file masking for coding agents (Codex, Claude Code, Cursor, Windsurf)
Local-first architecture (no telemetry, cloud-independent)
Self-hostable via Docker (single command)
Live dashboard with request stats (scan time, masked count, model, language)
Mask mode and Route mode (local LLM for sensitive requests)
Open-source and auditable (Apache 2.0 license, GitHub repo)
Cross-platform (macOS, Linux, Windows via Docker)
Keyboard shortcut (⌘⇧K) to mask selected text manually
AI-native enriched classification with 95%+ precision
Data Security Posture Management (DSPM)
Omni Data Loss Prevention (DLP) with unified AI brain
AI Guardian for AI activity tracking and governance
AI Security Posture Management (AI SPM) for shadow AI discovery
Access Trail for auditing human and AI data access
Browser Shield for browser-based data protection
AI Runtime Protection for real-time AI security
AI Protect to prevent sensitive data leakage during AI use
Automated remediation via Tines and Torq integrations
Data Subject Request automation (add-on)
DataWatcher managed services (add-on)
Compliance mapping to PCI DSS, DORA, GDPR, CCPA, HIPAA
Identity-aware data security with user context
Real-time risk reduction prioritization and guided remediation
