AI Runtime & Agent Guardrails
Products that monitor and control artificial intelligence (AI) agents and workloads at runtime, enforcing guardrails on actions and tool use.
MCP Manager
MCP Manager, based in New York, is an AI security company focused on bringing consent and trust controls…
Operant AI
Operant AI secures cloud-native and AI workloads at runtime, with a particular focus on the Model Context Protocol…
Lasso Security
AI security platform discovering shadow AI usage and governing autonomous agent behavior with an 'Intent Security' model.
SplxAI
A Zagreb-founded startup that automates red-teaming for AI agents, running simulated attacks to catch prompt injection before deployment.
Prime Security
A Series A New York startup using agentic AI to catch security design flaws before code is written.
Eve Security
Austin startup building "agent-in-the-loop" runtime policy enforcement that restricts what AI agents are allowed to do based on…
Trustwise
Trustwise builds an AI runtime control plane that intercepts and evaluates agentic AI actions before execution, authorizing, blocking,…
Dreadnode
An offensive-security startup that red-teams AI models and agentic systems before and after deployment, founded by former NVIDIA…
Opsin Security
Opsin gives enterprises visibility and risk scoring for the AI tools and agents employees adopt on their own,…
AIR
AIR builds a runtime context firewall that vets and monitors the skills, plugins, and MCP tools enterprise AI…
Clawvisor
An open-core authorization gateway that scopes and audits what AI agents are allowed to do before granting them…
Adversa AI
Bootstrapped AI red-teaming and adversarial-vulnerability research firm offering continuous red-teaming for agentic AI and coding-agent security.
PVML
PVML provides a differential-privacy-based secure data access layer that lets AI agents and LLMs query enterprise databases via…
AI EdgeLabs
Autonomous, AI-native runtime security platform protecting AI workloads, LLMs, AI agents, and the infrastructure they run on.
Outerlimit
A pre-seed New York startup building a decentralized, cryptographic authorization layer that fragments credentials so no single compromised…
Velatir
Danish AI-governance platform adding human-in-the-loop review and control over enterprise AI agent actions.
Promptfoo
A San Francisco open-source LLM and agent security-testing tool, used by over 25% of the Fortune 500, acquired…
Credal AI
Credal (Y Combinator W23) is a governance and security layer for enterprise AI agents — scoped tool access,…
Tynapse
A Seoul-based seed-stage startup building a runtime "AI Trust Layer" to detect hallucinations, data leaks, and jailbreaks in…
Arrakis Security
Autonomous workforce governance platform that discovers AI agents, monitors their behavior, and blocks policy violations across SaaS, cloud,…
Eve Security
Austin-based startup building EveGuard, a runtime observability and policy-enforcement layer for securing autonomous AI agents in the enterprise.
OpenAI
A leading AI lab that acquired AI red-teaming tool Promptfoo in March 2026 to embed security testing in…
Manufact
Manufact is a YC-backed startup building open-source SDK and cloud hosting infrastructure for deploying MCP servers that connect…
Tumeryk
AI trust scoring, red teaming and runtime guardrails for large language models and AI agents.