Privacy rules that actually work
Define what AI can and can't know about you. Automatic redaction, sensitivity classification, and compliance templates for regulated industries.
Governance at every layer
The policy engine evaluates every piece of context before it reaches any AI platform.
Redaction Rules
Pattern-based and semantic redaction. Strip company names, financial data, health info, or any custom pattern before compilation.
Sensitivity Classification
Tag data as public, internal, confidential, or restricted. Each level controls who and what can access it.
Retention Policies
Set time-to-live on state entries. Stale context auto-expires. Configure per-field or globally.
Data Residency
Control where your identity is stored and processed. Specify allowed regions for compliance.
Conflict Resolution
When policies overlap, most-restrictive-wins. Deterministic, predictable, auditable.
Audit Trail
Every policy evaluation logged with actor, action, outcome, and timestamp. Export for compliance.
Pre-built for your industry
Start with a compliance template and customize. Or build from scratch.
Healthcare
PHI pattern detection, automatic redaction of patient identifiers (names, MRNs, DOBs), configurable retention windows, healthcare-specific domain scoping.
EU Data Protection
PII detection and redaction, data residency controls, consent management, right-to-erasure support, purpose limitation enforcement.
Payment Data
Card number (Luhn-valid) redaction, CVV/CVC detection, scoped processing domains, 1-year retention, payment-specific sensitivity classification.
Build Your Own
Define custom patterns, sensitivity levels, retention rules, and residency constraints. Full YAML configuration with validation.