Product
Meridian Assist Meridian Runtime Meridian Console
Signals
On live calls On AI agents
For business Pricing FAQ Book a demo →
Meridian is Lurion's real-time social engineering defense platform for voice communication. It detects manipulation during live calls and tells employees or AI agents what to do before they comply with a dangerous request.
Training helps before the call. Meridian helps during the call — when attackers apply pressure and employees are expected to remember policies from memory. It's live defense at the moment of decision, not a replacement for training.
No. Meridian is consent-first. Employees activate it during suspicious or uncertain calls. Audio is processed on our servers during active sessions to detect social engineering patterns. Record only if you are authorized under the laws in your jurisdiction.
Meridian Assist is the employee-facing app for live-call defense. Meridian Runtime is the API/SDK for AI voice agents and call infrastructure — it returns policy decisions (allow, verify, warn, escalate, block). Meridian Console is the admin layer for incidents, policy, team management, and audit exports.
Yes. Meridian Runtime integrates with AI voice-agent platforms to catch trick questions, role changes, and attempts to pull private information — returning a clear allow, verify, or block decision before the agent acts.
STIR/SHAKEN helps authenticate caller ID routing. Meridian evaluates whether the interaction itself is safe to comply with. Caller ID can be authenticated while the caller is still manipulative, unauthorized, or socially engineering the target.
Security, fraud, and operations teams whose staff answer high-risk inbound and outbound calls — healthcare front desks, IT help desks, financial support, executive assistants, and organizations deploying AI voice receptionists.
Book a demo to see Meridian detect social engineering live, or email founders@lurion.org to discuss a design-partner pilot.

Still have questions?

We're happy to walk through your specific call workflows and security requirements.