Platform Overview
How Mirror products fit together across application, agent, identity, policy, and validation workflows
Platform Overview
Mirror products can be used independently or as a composed stack.
Product layers
Mirror VectaX
Encrypted AI application surface for gateway-backed AI, memory, vectors, FPE, PRE, RBAC, ECAD, policy, and FHE workflows.
Code Prism
Coding-assistant and secure development workflows built around gateway, policy, and protected agent execution.
AgentIQ Policy Engine
Prompt, tool, retrieval, and output policy enforcement for assistants and agents.
AgentIQ Identity
Agent registry, capability tokens, delegated execution, and MCP/server protection.
Agent Extensions
Connector readiness, MCP server onboarding, tool inventory, and skill attachment.
Mirror Behavioral Defense
Human and AI-agent activity coverage across endpoints, browsers, sanctioned Gateway workflows, and policy evidence.
DiscoveR
Red-teaming, browser scanning, streaming API scanning, and CI-driven adversarial validation.
How they compose
Mirror VectaX handles gateway-backed AI, encrypted memory, vectors, and protected content workflows.
AgentIQ Policy Engine decides what prompts, outputs, retrieval results, and tool calls are allowed.
AgentIQ Identity issues scoped tokens, owns the agent registry, and preserves delegation lineage for agents and tools.
Agent Extensions manages connector readiness, MCP servers, tools, and skills before agents can use external actions.
Behavioral Defense connects endpoint, browser, agent, Gateway, and policy evidence into cases and response workflows.
DiscoveR validates the resulting assistant and API behavior with red-teaming and adversarial scans.
Use it two ways
Product-by-product
Start with the one product that matches your immediate need.
Composed stack
Use the products together when you need secure agents with identity, policy, protected data paths, and validation.
The fastest path for new teams is usually: start with Mirror VectaX, add AgentIQ Policy for runtime controls, add AgentIQ Identity when agent identity matters, add Agent Extensions for tools and MCP onboarding, add Behavioral Defense for endpoint and AI-agent activity coverage, then add DiscoveR and Governance for validation and evidence.