Protecting GenAI Applications in Real-time

Clestiq Shield

The Unified AI Security Platform. Secure your LLMs with defense-in-depth architecture, from advanced input sanitization to output hallucination checks and PII redaction.

Defense-in-Depth Architecture

Requests flow through our intelligent security layers—Sentinel and Guardian—ensuring safety before and after every LLM interaction. Managed by Eagle-Eye and observed via Datadog.

Select a simulation to visualize the architecture
Client
Redis Cache
Gateway
Eagle-Eye
PostgreSQL
Sentinel
LLM PlatformProvider Agnostic
Guardian
Datadog AgentLogs, Traces, Metrics
Kube AgentInfra Stats

Datadog Observability Platform

Ingesting Metrics, Logs, & Traces

Monitors & Detectors
Email Alerts
Case Management

Core Services

Comprehensive Protection Suite

Four specialized microservices working in harmony to secure your AI infrastructure.

Sentinel

The Input Security Agent. Specialized in detecting and neutralizing incoming threats.

  • SQL & Command Injection
  • XSS Protection
  • PII Redaction
  • Input Sanitization
  • LLM Forwarding

Guardian

The Output Validation Agent. Ensures AI generated content is safe, accurate, and on-brand.

  • Hallucination Detection
  • Toxicity & Tone Check
  • Data Leak Prevention
  • PII Leak Detection
  • False Refusal Check
  • Structured Output

Eagle-Eye

Identity & Management Layer. Complete visibility and control over your API usage.

  • Usage Analytics
  • User Management
  • App Management
  • API Key Management

Gateway

The central nervous system orchestrating secure communication.

  • Ultra-low Latency
  • Request Routing
  • Load Balancing
  • Traceability
  • Rate Limiting (Redis)
Powered by Datadog

Full Visibility with Datadog

We don't just secure your AI; we give you X-Ray vision into it. Clestiq Shield integrates natively with the Datadog Cloud Platform to provide end-to-end observability.

APM Traces

End-to-end request tracing from Gateway to LLM models.

Unified Logs

Centralized logging for all 4 microservices with context.

Real-time Metrics

Track token usage, latency, and threat detection rates.

K8s Cluster Stats

Infrastructure visibility via Kube State Metrics integration.

Monitoring Capabilities

Continuous Profiler
Infra Monitoring
APM
Log Management
Metrics
Traces Correlation
Dashboards
Monitors
Alerting
Case Management