TRUST CENTER

Security at the Speed
of Thought.

Production-grade governance. Privacy-first architecture. SOC2 aligned.

Not a black box. Every action your employees take is logged, explained, and auditable.

The Isolation Architecture

Your data flows through a Zero Trust pipeline. We process your instructions. We never train on your secrets.

Your Environment

AES-256 encrypted at rest

Secure Origin

Zero Trust Gate

RBAC + Risk Scoring

Policy Engine
CORTEX ENGINE

Cortex Engine

180+ Cognitive Methodologies

Processing

Action Output

Results Only. Context Expires Automatically.

Delivered
"Your data never trains our models. By design."

LiquidCortex creates isolated processing contexts for every task. Once the action is complete, context expires automatically. We couldn't sell your data even if we wanted to.

Security Pillars

Hardened protection at every layer

Envelope Encryption

Your data is encrypted with your own unique key. Even our systems can't read your plaintext without your authorization.

AES-256 + PBKDF2 · TLS 1.3 in transit

Zero Trust Architecture

Every request is verified. Every action is logged. No implicit trust. RBAC policies define exactly what each employee can access.

NIST SP 800-207 Aligned

No Model Training

Your prompts and data are never used to train AI models. We process. We don't learn from your proprietary information.

Zero Training Policy

Auto-Expiring Context

Cached data self-destructs. Configurable TTL ensures your context doesn't persist beyond your session requirements.

7-Day Default TTL

Jailbreak Shield

Prompt injection attacks are blocked at the gate. Our security middleware detects and neutralizes adversarial inputs.

Real-time Detection

Decision Trace™

Every AI decision comes with a plain-English explanation of why it was made. Click any action to see the full reasoning chain — who triggered it, what data was considered, and what alternatives were evaluated. Every reasoning chain is recorded and queryable.

Clickable Reasoning Trails
The Hidden Risk

Stop Shadow AI

Your employees are already using unauthorized AI tools with company data. No audit logs. No credential management. No idea what data is touching what API.

The Problem: Shadow AI

  • • Developers running AI agents on personal servers
  • • Sales reps feeding client data into random AI tools
  • • No visibility into what data is being processed
  • • No way to shut it down when something goes wrong
  • • Zero compliance audit trail

The Solution: LiquidCortex

  • • Managed AI employees with scoped permissions
  • • Every action logged via Decision Trace™
  • • Per-employee NCU budgets control spending
  • • Kill switch halts any employee instantly
  • • Full audit trail for compliance

Instead of banning AI, give your team the managed alternative. Same productivity, full control.

The Chain of Command

You're the CEO. Every AI employee reports to you.

1

Granular RBAC

Define exactly what Atlas can see versus what Hunter can scrape. You're the CEO—they follow your clearance levels.

2

Resource Boundaries

Each employee operates within strict resource limits. No employee can access data outside their designated scope.

3

Kill Switch

Instant termination of any employee or workflow. Full control remains with you at all times.

Compliance Frameworks

Built for regulated industries. We are pre-certification — the platform is designed against these frameworks; formal audits are planned post-launch.

SOC 2

SOC 2 Trust Service Criteria

Principles-Based

Our architecture is designed against SOC 2 Trust Service Criteria. We are pursuing formal certification.

GDPR

GDPR Principles

Privacy-First

Our data practices align with GDPR principles. See our Privacy Policy for EU handling details.

HIPAA

HIPAA

BAA on request

HIPAA-ready architecture. Business Associate Agreements available on request for covered entities.

ISO

ISO 27001

In Progress

Working towards ISO 27001 certification for information security management.

Need a security review?

Request our security questionnaire, penetration test summary, or schedule a review call with the security team.

Contact Security Team

Our commitments, in four numbers

SOC 2
Built to Trust Criteria
GDPR
Privacy-First
256-bit
AES envelope encryption
0
Models trained on your data