PRIVACY GATEWAY & TRANSPARENCY
AI Usage & Redaction Audit Trail
Complete audit trail of all external LLM prompts, model identifiers, and privacy redaction counts.
Verity Zero-Leakage Privacy Gateway
Before user content is sent to Groq AI, Verity strips out emails, phone numbers, and confidential codenames (like Project Falcon), replacing them with synthetic tokens (`[REDACTED-CONFIDENTIAL-1]`). The mapping is preserved securely in memory to restore original phrases in suggested fixes.
NO RAW PII SENT TO LLMs PROMPT-INJECTION SAFE DELIMITERS AUDITABLE PER-REQUEST LOGS
Prompt Dispatches (0)
Live Audit TrailLoading privacy audit logs...