tools
Noma AI Security Platform
Noma Security helps enterprises discover, govern, test, and protect AI applications, agents, models, MCP servers, and tools.
Noma Security is an enterprise AI security and governance platform. It maps AI assets and risks, controls agent access, tests AI applications and agents for attacks, and detects risky behavior during runtime.
Security, engineering, compliance, and business teams use it across homegrown AI, SaaS agents, and endpoint agents. The site does not list self-serve plans or public pricing; buyers are directed to request a demo.
Features
- Discovers AI agents, models, MCP servers, skills, and tools across environments
- Maps agent connections, permissions, data access, and security risks
- Defines and enforces access policies for agents, tools, and MCP servers
- Assigns agent identities and supports approved, review, or blocked states
- Tests AI applications and agents for prompt injection, jailbreaks, and data leakage
- Detects prompt injection, data exfiltration, scope violations, and rogue behavior
- Maps findings to NIST AI RMF, EU AI Act, ISO 42001, OWASP, and MITRE ATLAS
Use cases
- Inventory AI agents and assets that security teams did not know were deployed
- Control which agents, MCP servers, skills, and tools employees can use
- Test customer-facing AI applications for multi-turn attacks
- Detect data exfiltration and behavior that drifts from an agent's intended task
- Produce AI security, compliance, test-result, and audit reports