USE CASE / DATA PROTECTION

Stop sensitive data before it crosses an AI boundary.

Apply configurable PII, secret, and denylist detectors to prompts and model responses at the on-prem gateway, with policies scoped by key, project, team, or organization.

CONTROL OUTCOMES

Move from fragmented access to enforceable policy.

01

Pre-call inspection

Inspect text inputs before the selected provider receives them, including text parts inside multimodal requests.

02

Post-call inspection

Evaluate model output before it is returned to the calling application.

03

Policy actions

Configure allow, flag, redact, or block behavior by detector, direction, and governance scope.

04

Audit-ready evidence

Persist guardrail decisions with the virtual key, provider, model, usage, and final request outcome.

REFERENCE ARCHITECTURE

Controls execute in the request path.

01

Authenticate the key

Resolve the virtual key and its organization, team, project, and applicable policy scopes.

02

Inspect the request

Run configured PII, secret, and denylist detectors before routing upstream.

03

Route and inspect

Call an approved deployment, then apply response-direction guardrails before returning output.

04

Record the decision

Write the call and guardrail events to the durable audit record and optional SIEM projection.

GATEWAY TELEMETRY
ILLUSTRATIVE REFERENCE VIEW
REQUESTS PROXIED18.4k24H WINDOW
VIRTUAL KEYS42ORG / TEAM / PROJECT
GUARDRAIL BLOCKS37PII / SECRET / DENYLIST
GATEWAY P95312msSLO PASS
TIMEVIRTUAL KEYMODEL ALIASPROVIDEROUTCOME
14:32:08vk_finance_07fastanthropicSUCCESS
14:31:55vk_support_02reasoningopenaiFALLBACK
14:31:41vk_research_12summarizegeminiBLOCK

TRUST ARCHITECTURE

Built for security review, not security theater.

Map virtual-key access, encrypted provider credentials, guardrail decisions, and audit evidence to the controls your enterprise already operates.

SOC
SOC 2 control alignment

Architecture supports evidence collection for access, change, and monitoring controls.

ISO
ISO 27001 readiness

Map credential, access, policy, and audit practices to ISMS control objectives.

AI
NIST AI RMF mapping

Operational visibility across governed model access, guardrails, usage, and provider outcomes.

Control-alignment statements describe product architecture and are not claims of EnvisionAI certification.

WORKS WITHOpenAIAnthropicGeminiVertex AIGroqOllama

TECHNICAL EVALUATION

Review the control path with your security and platform teams.

Map the architecture to your providers, application clients, governance scopes, data boundaries, and operating requirements.

ENTERPRISE INQUIRY / SECURE FORM

Request an enterprise demo

Tell us about your environment. We will route your request to a technical specialist.

Business inquiries only. No free-tier or consumer accounts.