01
Endpoint Outside the Zone
The request starts at the edge
AI agent requests originate from user endpoints that may operate outside the security boundaries protecting critical enterprise systems.
X-PHY AI ARMOR
Hardware Rooted Protection for AI Agents and Enterprise Resources
X-PHY AI Armor protects the endpoint where every AI agent request begins and helps safeguard the enterprise resources connected through MCP servers.
Combining X-PHY Cybersecure SSD, AI Armor Agent and AI Armor Platform, it brings hardware rooted trust, authenticated integration, security policies and rules into the AI workflow.
X-PHY Cybersecure SSD · AI Armor Agent · AI Armor Platform · Hardware Root of Trust
THE ENDPOINT TRUST GAP
Every AI Agent Prompt Starts at an Endpoint That May Already Be Compromised
AI agents connect users to sensitive business systems and enterprise data. However, authenticated sessions alone cannot guarantee endpoint security. A compromised device may expose trusted access to attackers, putting connected resources at risk.
X-PHY AI Armor helps close this gap by verifying trust at the hardware level, not just at login.
02
Session Token Authentication
Authentication backed by hardware trust
Session tokens establish authenticated access, while X-PHY Hardware Root of Trust provides an additional hardware backed foundation for endpoint trust.
03
Connected AI Workflows
One connection can reach multiple systems
Agent Servers and connection services route AI requests to enterprise resources, making the trust and security of each connected endpoint essential.
04
Critical Resources at Risk
Protection must extend beyond the endpoint
The systems an AI agent can reach, not the device itself, are the real targets behind every request.
ENDPOINT→SESSION AUTHENTICATION→CONNECTION SERVER→ENTERPRISE RESOURCES
X-PHY PROTECTION STACK
Five Components Behind X-PHY AI Armor
01
X-PHY Cybersecure SSD
Hardware security at the endpoint
Establishes a hardware rooted security foundation within endpoint storage, helping protect against compromise and unauthorized activity.
02
AI Armor Agent (DC Agent)
Connects hardware security to the platform
Communicates SSD status and security events to the AI Armor Platform and executes supported security response commands.
03
AI Armor Platform Web API
One authenticated integration point
Provides an authenticated interface for endpoint communication, security event collection and integration with AI agent workflows.
04
Authentication Layer
Hardware backed command authentication
Uses API key authentication together with the SSD's unique secret key and a one time password model to verify supported security commands.
05
Command Dispatcher
Centralized security response
Coordinates Lock, Wipe and status query commands for enrolled endpoints through the AI Armor Platform.
PROTECTION & INTEGRATION
Independent Protection. Connected Response.
X-PHY AI Armor combines autonomous hardware security with authenticated platform integration.
Firmware protection and embedded hardware sensors operate independently at the endpoint, while the AI Armor Platform connects endpoint security events and response capabilities to the wider AI agent workflow.
LOCAL / INDEPENDENT
Security That Starts in Hardware
- Firmware level protection
- Embedded hardware sensors
- Endpoint security foundation
CONNECTED / MANAGED
Security Response Across AI Workflows
- Authenticated API integration
- AI Armor Platform
- AI agent response coordination
AI AGENT ARCHITECTURE
How AI Agents Connect to Enterprise Resources
AI agent workflows rely on authenticated sessions and connection services to reach enterprise systems. The MCP Server is the common link to those resources, while AI Armor adds hardware rooted security around the workflow.
REQUEST
User / Application
X-PHY SECURITY
Endpoint
AUTHENTICATED
Agent Server
STANDARD CONNECTION
MCP Server
CONNECTED RESOURCES
Databases
LLMs
Business Tools
AI Services
RESULT RETURNS ← MCP SERVER ← AGENT SERVER ← ENDPOINT
01
User Prompt
A user or application initiates an AI agent request from the endpoint.
02
Session Authentication
The Agent Server authenticates the session before processing the request.
03
Request Routing
The AI workflow passes the request toward the appropriate MCP Server.
04
Resource Access
The MCP Server connects the request to the relevant database, AI model, business tool or service.
05
Response Returns
The result travels back through the AI workflow to the requesting endpoint.
BEYOND ENDPOINT PROTECTION
Hardware Rooted Trust for Everything Behind the MCP Server
Everything an AI agent can reach through an MCP server becomes part of the AI ecosystem, and part of its attack surface.
X-PHY AI Armor extends protection to these connected resources by combining three security elements into one foundation.
The result is stronger trust in every interaction between endpoints, AI agents and enterprise systems.
X-PHY AI ARMOR SECURITY FOUNDATION
Hardware Root of Trust + Security Policies + Rules
Hardware Root of Trust
Trust Begins in Hardware
The Cybersecure SSD gives each endpoint a hardware backed starting point for trusted authentication across AI connected environments.
Security Policies
Protection Guided by Defined Policies
Policies establish requirements for trusted operations and help align AI agent interactions with enterprise security expectations.
Security Rules
Consistent Security Decisions
Security rules help guide the handling of security events and supported responses, strengthening protection across connected AI workflows.
WHAT THE FOUNDATION HELPS PROTECT
Business Data
Records that AI agents query on a user's behalf
AI Models
The models and inference services your agents rely on
Workflow Tools
Applications agents can act through, not just read from
Digital Assets
Intellectual property and sensitive information across the business
Secure the Endpoint. Strengthen Trust. Protect Connected Resources.
From the SSD at the endpoint to the resources behind the MCP Server, AI Armor helps build trust into each step of the AI workflow.
WHEN TRUST IS HIJACKED
The Same Architecture, Seen by an Attacker
The workflow itself does not change. When attackers gain control of an authenticated session, they may misuse the trusted AI workflow to reach connected business systems.
COMPROMISED
Endpoint
AUTHENTICATED
Agent Server
STANDARD CONNECTION
MCP Server
WHAT AN ATTACKER MAY ATTEMPT
Data Theft
Model Misuse
Tool Abuse
Service Disruption
SAME ROUTE · SAME SESSION · DIFFERENT OPERATOR
01
Endpoint Compromise
The attack begins at the device
Malware, phishing or stolen credentials may give attackers access to an endpoint.
02
Session Hijacking
Attackers inherit trusted access
A compromised session may allow unauthorized activity under legitimate user permissions.
03
Trusted Channel Abuse
Malicious requests follow legitimate paths
Attackers may attempt to use existing AI connections to reach internal resources.
04
Enterprise Exposure
Exposure spreads across connected systems
Every system connected to the compromised workflow may face increased security risk.
BLAST RADIUS
One compromised endpoint can expose multiple connected resources, making hardware rooted protection and policy driven security essential.
BUILT INTO THE ENDPOINT
Protection That Starts Inside the SSD
Autonomous Firmware Detection
AI One Core monitors storage activity at firmware level to identify suspicious behavior and potential threats, including previously unknown attack patterns.
AES-256 XTS Encryption
Hardware based encryption protects data stored on the X-PHY Cybersecure SSD.
Embedded Hardware Sensors
Built in sensors support physical tamper detection and security monitoring independently of installed software or network connectivity.
Automatic Lockdown
Supports automatic security lockdown behavior when configured drive removal conditions are detected.
CLEAR ANSWERS. ZERO CONFUSION.
Frequently Asked Questions
What is X-PHY AI Armor?
It is X-PHY's security solution for AI agent workflows, built on three parts: the Cybersecure SSD inside the endpoint, the AI Armor Agent running on the device and the AI Armor Platform for central management.
What does AI Armor protect beyond the endpoint?
Everything AI agents reach through MCP servers, from company data and AI models to the business tools and assets behind them.
How does Hardware Root of Trust strengthen AI security?
Session credentials can be stolen from a compromised device. Hardware Root of Trust adds a second anchor inside the SSD itself, which is far harder for an attacker to take over.
What role do security policies and rules play?
Hardware trust tells AI Armor whether an endpoint can be trusted. Policies and rules decide what happens next, so the same situation always gets the same response.
What does the MCP Server do?
The MCP Server provides a standardized interface that allows AI applications and agents to interact with supported enterprise tools, data sources and services.
What does the Agent Server do?
The Agent Server manages authenticated interactions and routes requests within the AI agent workflow.
How does the AI Armor Platform connect to endpoints?
Through the AI Armor Agent on each endpoint. The agent reports to the platform over an authenticated API and carries out the commands the platform sends back.
How are security commands authenticated?
Each command is checked twice: against a valid API key, and against a one time password tied to that SSD's unique secret key.
What security response actions are supported?
Three commands for enrolled endpoints: Lock to secure the drive, Wipe to erase its data and status query to check its current state.
Does hardware protection require an internet connection?
No. The protection runs locally on the drive, so it keeps working even when the endpoint is offline.
SECURE YOUR AI ECOSYSTEM
Protect the Endpoint. Safeguard Everything Behind It.
See how AI Armor protects the device where AI agent requests begin and strengthens trust across every system those requests reach, in a live session with the X-PHY team.
Book A Demo SessionAI Agent Security · Endpoint Protection · Device Trust
