Noma Security

Noma Security

Noma Security is an agent security and governance platform for enterprises adopting AI agents. Noma finds every agent, MCP server, skill and tool running across endpoints, SaaS platforms and cloud environments, lets security teams decide what each one may access, and watches agent actions at runtime. When a sequence of individually permitted actions adds up to something risky, Noma can block the action, steer the agent, mask data, alert a person or trigger a kill switch. It also red-teams AI applications with multi-turn attacks.

Production credibility: Noma says more than 10% of the Fortune 100 use its platform and that Gartner named it a Market Shaper in the 2026 Emerging Market Quadrant for AI Application Security, start-up vendors; both are the company's own statements. Its leadership listed on the site includes co-founders Niv Braun (CEO) and Alon Tron (CTO) and CISO Diana Kelley. The Noma Labs research team publishes agent vulnerability findings, including SneakPeek, a data-exfiltration issue in the Claude Code desktop preview pane. Noma's products are AI-SPM for discovery and posture, AI-DR for runtime detection and response, Agent Access Control and AI Red Teaming, and its Open Enforcement approach applies one policy at the control points a company already runs. Noma does not publish pricing; access is through a sales demo.

Key Features

• Discovery of AI agents, MCP servers, skills and tools across endpoint, SaaS and cloud
• Approve, review or block each agent, MCP server and skill, scoped by user or team
• Runtime monitoring that judges each agent action against the whole session
• Responses that block, steer, mask data, alert a human or kill an agent
• Multi-turn AI red teaming and agent supply chain risk checks
• Compliance support for AI regulation across the AI environment

Ideal Use Case

Security teams at large companies whose employees and engineers are adopting agents faster than policy can keep up: coding agents on laptops, agents built on SaaS platforms, and agents engineers deploy in the cloud. Noma gives them one inventory and one policy layer across all three instead of a separate control in each platform.

How Noma Security differentiates

Zenity is the closest rival, with a similar focus on discovering and governing agents across SaaS and endpoints. HiddenLayer and Mindgard started from securing ML models and red-teaming them, whereas Noma's centre of gravity is agent behaviour at runtime and agent access. Noma covers endpoint, SaaS and cloud agents in one platform; it is an enterprise product with no self-serve tier.

FAQ

Q: What is Noma Security? A: Noma Security is an agent security and governance platform. It discovers AI agents, MCP servers and tools across an organisation, controls what they can access, and monitors and stops risky agent behaviour at runtime.

Q: How much does Noma Security cost? A: Noma does not publish pricing. You book a demo with its sales team to get a quote.

Q: Does Noma Security cover MCP servers? A: Yes. Noma discovers MCP servers, skills and tools alongside agents, and lets security teams approve, review or block each one.

Q: What environments does Noma Security support? A: Noma covers agents on employee endpoints, agents running on SaaS platforms, and agents engineers build and deploy in the cloud.

Q: How does Noma Security compare with Zenity? A: Both discover and govern AI agents across SaaS and endpoints. Noma emphasises runtime context, weighing each action against the full session, and adds AI red teaming in the same platform.

tl;dr

Noma Security is an enterprise platform that discovers AI agents and MCP servers, controls their access and stops risky agent actions at runtime across endpoint, SaaS and cloud. It says over 10% of the Fortune 100 use it; pricing is by quote.