One agent. Your network. Nothing else in the loop.
The xRazor agent pairs a private on-prem LLM with industry-proven scanners. The core ships today; two modules are on the bench — both extend the same agent, the same boundary, and the same audit trail.
Know every door before someone else does
ASM keeps a continuous inventory of everything the agent could reach — external and internal, documented and forgotten. The map updates itself as your environment changes, and everything it finds feeds the scan queue automatically.
Feeds the autonomous scan queue — every new asset is a new target on the next edition.
Get notified when it ships- External and internal asset inventory, continuously refreshed
- Change tracking between scan editions — new services surface as changes, not surprises
- Findings feed the autonomous scan queue without manual triage
- The same boundary discipline: discovery runs entirely inside your network
An agent that plans its own attack path
Today's agent scans what you point it at. The autonomous module lets it plan: chain findings the way a human tester would, follow the attack path across services, and re-test after every deployment — without a prompt from you.
Extends the scan engine — same agent, same boundary, same audit trail.
Get notified when it ships- Self-planned attack chains across services, not just per-target scans
- Re-test on every commit or deployment — findings go stale less often
- Full reasoning trail per finding: why the agent tried what it tried
- Runs on the same on-prem LLM — no external calls, ever
What the agent already does
The core is deployed and running inside customer networks today. Everything above builds on it — nothing replaces it.
Asset discovery
Enumerate hosts, endpoints, and APIs across networks and applications
AI-led attack simulation
Contextual attack paths orchestrated by the on-prem LLM
Hybrid scanning engine
LLM-guided simulation combined with industry-proven scanners
Correlation & deduplication
Signals merged into ranked findings — precision over noise
Reports & remediation
Actionable reports with remediation context, pushed into your tools
Full audit trail
Every scan, finding, and action logged for compliance
Runs where you run it
Three deployment shapes, one agent. Licensing follows the environment you deploy against.
Data center
Standard on-prem deployment inside your controlled environment
Private cloud
Same agent, your cloud account, your network policy
Air-gapped
No upstream at all — offline install, offline operation
Bring a target. Keep your network.
The fastest way to evaluate the platform is to watch it run against something real — inside your boundary, with your policies, on your hardware.