PII Shield strips names, SSNs, emails, addresses, phone numbers, account numbers, and dates of birth before requests leave our servers. The model sees placeholders. Your screen still shows the real names. No AI training on your data — per Google's paid Gemini API terms in the default configuration. Per-firm tenant isolation. Full audit trail.
PII Shield redacts names, SSNs, emails, addresses, phone numbers, account numbers, and dates of birth before any AI model call
Your screen still shows real names — the model only sees placeholders like [NAME_1] or [SSN_1]
Enterprise Gemini API: no AI training on your data — brief abuse-detection only, separate from training
Per-firm tenant isolation — a leaked credential cannot read another firm's matters
AES-256 encryption at rest, TLS 1.3 in transit, signed object-storage URLs expire in 60 minutes
Append-only audit log of every read, write, AI call, and login — exportable as CSV