Application policies
Save input/output rules and assign an immutable policy version to each application. Allow, mask, redact, tokenize or block detected sensitive data.
PRIVACY PLATFORM
Controls for the AI traffic you connect to PrivoNest. Start with the integration walkthrough.
Save input/output rules and assign an immutable policy version to each application. Allow, mask, redact, tokenize or block detected sensitive data.
Inspect non-streaming text requests and responses for operator-configured model providers. Bring your own provider key. Files, images and tool calls are not supported by the gateway.
Separate protect, gateway, test and reveal permissions. PV2 tokens bind company, app, optional subject and expiry. Backend end-user authorization remains required.
Inspect policy decisions, detected-type counts, credits, errors and latency. Raw prompts, documents, token values and provider keys are excluded from audit records.
Run synthetic regional examples and authorized custom model/target probes. Track exact canary results and matching-fixture comparisons. These checks are not a complete security assessment.
Compare imported labeled entity spans using precision, recall and F1. The tool scores supplied predictions; it does not independently verify competitors.
Maintain app inventory, vendor reviews, risks and regional review evidence. Export JSON reports. Evidence coverage is not legal certification or a security score.
The original recognizers and document/OCR paths remain available. Configured language/country coverage is exposed by /v1/capabilities; accuracy varies by data type and configuration.
Owner, admin, developer and viewer roles; verified existing teammates; per-app keys; retention settings and daily maintenance support.