8 integrations
In this group
LiveOpenTelemetry (OTLP/HTTP)Turns the AI telemetry your applications already produce into a security view of models, agents and tools.View integrationLiveLLM proxyGives security teams one control point for model calls, with access checks, usage counts and a stop.View integrationLiveMCP GatewayPuts a control point between AI agents and tool servers, so risky calls can be held or blocked.View integrationLiveDatadogPlaces AI security events beside the monitoring your engineers and on-call teams already use in Datadog.View integrationLiveLangSmithGives security teams a view of what LangChain and LangGraph agents do, using LangSmith data engineers already collect.View integrationLiveLangfuseTurns Langfuse traces, sessions and evaluations into a security view of the LLM applications your teams run.View integrationLiveArizeTurns Arize evaluation results into a security signal, showing when an agent drifts from its usual behavior.View integrationLiveAgent-to-Agent (A2A) protocolApplies tool-call controls to agent-to-agent traffic, so one agent's request to another can be held or blocked.View integration