Skip to main content
As organizations adopt AI systems, they must safeguard their AI models, applications, and data during live operations. Organizations must continuously monitor and protect against threats that occur during live interactions, like prompt injection, data leakage, and model misuse. HiddenLayer’s Runtime Security helps organizations:
  • Detect AI attacks in real time before they cause damage
  • Respond faster with predefined playbooks and expert guidance
  • Minimize downtime and business disruption from AI incidents
AI Runtime Security is HiddenLayer’s real-time protection for LLM applications. It monitors model inputs and outputs and can detect, redact, or block content. It detects malicious input prompts and/or undesired output as they are sent to and returned from an LLM, and can (when configured appropriately) block content from being sent to the LLM or returned to the user. It has different modes of operation which can be flexibly employed, depending on the architecture already in place and the desired level of integration. HiddenLayer’s target operating model is designed to provide maximum flexibility, security, and operational independence for our customers. Our software is available as container images, allowing for seamless deployment, scaling, and integration into existing customer architecture. We provide pre-packaged, production-ready container images, which the customer deploys, configures, and operates independently within its own cloud or self-hosted Kubernetes infrastructure. This makes deployment and integration into an existing containerized infrastructure straightforward for DevOps teams.
RecommendedAgentic Runtime Security is the recommended way to protect AI applications. It is built for agentic AI (multi-turn sessions, tool calls, multiple providers) and reconstructs an agent’s many calls into one replayable session.
Start with Agentic Runtime Security for new work; its overview explains how it compares to the prior version. The prior version, AI Detection & Response (AIDR), remains fully supported. Console views for detections, interactions, projects, and policy are covered under Console. Runtime Security must be deployed separately from the Console—see Deploy AI Runtime Security using Helm. The LLM Sandbox, while part of the Console, is designed for testing scenarios and viewing the test results.
Release NotesRead the latest Runtime Security release notes.