Skip to main content
Post-Quantum · AI Agent Defense

HYVE Raptor

Post-Quantum ML-KEMAI Behavioral AnalysisZero Data AccessAir-Gap ReadyAlpha AI AdvisorElectron Desktop

A post-quantum AI agent defense platform with a full operator desktop, active defense commands, six-framework compliance engine, and an embedded offline AI advisor. Deploy on your infrastructure. Own every byte of threat intelligence.

By Anthony S. Owens · Vibe Software Solutions

🦅 Purchase & DeploySee the Command Center ↓
Architecture

Three Components. One Platform.

You run the infrastructure. Your clients run the shield. Nothing touches a third-party server.

🔑1

HYVE Backend

License Authority

Issues cryptographically signed bearer tokens to raptor-shield clients. Validates license keys you generate. Runs on your server — no third-party authority.

🖥2

Raptor Command

Operator Desktop + Threat Hub

Full Electron desktop application. Threat feed, shield management, defense command, multi-framework compliance posture, and Alpha AI Advisor — all in one operator workstation. ML-KEM-768 encrypted inbound channel on port 8443.

🦅3

Raptor Shield

AI Agent Watchdog

Runs on your client's machine. Monitors every AI agent for behavioral anomalies — capability drift, unexpected API calls, process escalation. Reports threats to your Command server in real time.

Deployment

Four Steps. Fully Operational.

⬇️
STEP 1

Run the installer

Double-click "1 - Install Server.exe". Both server components start automatically. Ollama and the Alpha AI model install in the background — no manual setup.

🖥
STEP 2

Open the Command Center

Launch the Raptor Command desktop app. Your full operator workstation opens: threat feed, shields, compliance posture, Alpha Advisor, and defense command — all live.

🗝
STEP 3

Generate a license key

Double-click "2 - Generate License Key.exe". A HYVE-XXXX-XXXX-XXXX key is created instantly. Give one key to each client — you are the key authority.

📦
STEP 4

Deploy raptor-shield to clients

Send clients the shield installer. They double-click, enter your server address and their license key, and the shield installs as a Windows service that starts monitoring immediately.

Competitive Advantage

What Nobody Else Offers.

Six capabilities absent from every other AI agent security product on the market today.

01
Harvest-Now-Decrypt-Later Protection

The only AI agent monitor with post-quantum encrypted telemetry.

Every other security tool sends threat data over standard encryption — which quantum computers will eventually break. Raptor protects threat intelligence at the application layer using post-quantum cryptography standardized by NIST. Nation-state adversaries recording your traffic today won't be able to decrypt it when quantum computers mature. No other AI agent monitoring platform addresses this threat model at all.

02
AI-Native Behavioral Analysis

Built specifically for AI agents — not repurposed from human endpoint tools.

Every EDR on the market was designed to monitor humans running software. AI agents behave differently — they operate autonomously, acquire new capabilities, call external APIs, and make decisions without human input. Raptor was engineered from the ground up to understand AI-specific behavior: what an agent is allowed to do, what it's actually doing, and what that delta means for your security posture.

03
Operator-Deployed Infrastructure

Your threat data never touches a third-party server. Ever.

CrowdStrike, Darktrace, Microsoft Sentinel — all SaaS. Your threat intelligence flows through their cloud. With Raptor, you run the infrastructure. Your clients' threat data flows to you and only you. Deploy in a classified facility, an air-gapped network, or a HIPAA-controlled environment. No Business Associate Agreement needed with a vendor. No FedRAMP authorization burden. You own the platform.

04
Self-Sovereign Licensing

You are the licensing authority. No third party can shut down your clients.

Every commercial security product has a licensing server in someone else's cloud. When it goes down, your clients go dark. Raptor's license authority runs on your server. You generate keys. You assign them. You revoke them. There is no HYVE licensing cloud. No outage, no vendor decision, and no pricing change can affect a single client deployment you control.

05
Zero Data Access Architecture

The operator cannot access client data — by design, not by policy.

Managed security services require operators to read client data. Raptor's architecture separates threat intelligence from client content at the protocol level. The operator receives behavioral signals — what changed, how severe, what action was taken — not the underlying data those agents were processing. A hospital's PHI cannot reach the operator. A law enforcement agency's case data cannot reach the vendor. The system enforces this. No policy document can override it.

06
Air-Gap Ready

Fully operational in air-gapped, classified, and offline environments.

Raptor was engineered to have zero external dependencies at runtime. No telemetry to HYVE. No update server to reach. No license check over the internet. Deploy in a SCIF, a classified government network, a disconnected industrial control environment, or any facility where outbound internet access is prohibited. No other AI agent security tool makes this claim.

Platform Capabilities

What HYVE Raptor Does.

🔬

AI Behavioral Baseline

Every AI agent is profiled at first contact — capabilities, API calls, network domains, process behavior. Deviations from baseline trigger scored threat events automatically.

⚛️

Post-Quantum Encryption

Shield-to-Command channel uses ML-KEM-768 (NIST FIPS 203) at the application layer. Threat data is quantum-safe in transit — not dependent on TLS alone.

🔐

Zero Data Access Architecture

Your Command server is cryptographically prevented from reading client plaintext data. You receive threat metadata, not client content. HIPAA and CJIS compliant by design.

🏔

Air-Gap Ready

All three binaries compile with CGO_ENABLED=0. Deploy in classified, air-gapped, or offline-only environments without dependency on external services.

📊

Threat Severity Scoring

Every threat event carries a severity score computed from behavioral delta, capability weight, and frequency. The command center surfaces highest-risk agents instantly.

🗝

Self-Sovereign Licensing

You generate license keys. You assign them. You revoke them. No licensing authority, no phone-home, no service dependency. HYVE-XXXX-XXXX-XXXX format, generated locally.

🤖

Alpha AI Defense Advisor

An embedded, fully offline AI advisor powered by a local LLM. Alpha has live access to your shield network, threat feed, and compliance state — and answers operational questions without sending data anywhere.

⚖️

Multi-Framework Compliance

Real-time posture scoring across CMMC 2.0, HIPAA, FedRAMP, CJIS, NIST 800-171, and SOC 2 simultaneously. Failing required controls surface instantly with evidence and remediation context.

🎮

Active Defense Commands

Issue ISOLATE, CONTAIN, NEUTRALIZE, or LOCKDOWN to any shield from the command center. Commands are encrypted and dispatched with a sub-100-millisecond delivery target; execution completes as fast as the target OS permits (Windows Firewall rule installation, process termination, or service stop).

🔄

Persistent Event Store

Raptor Command stores all threat events in a WAL-mode SQLite database with full restoration on restart. No data lost on process crash or server reboot.

Regulatory Alignment

Built for Regulated Industries.

Healthcare, defense, federal agencies, and law enforcement each have their own compliance requirements. Raptor is engineered to satisfy them all.

HIPAA

Zero data access model prevents PHI exposure at the infrastructure layer

CMMC Level 2

Air-gap deployment + post-quantum channel satisfies CUI protection requirements

FedRAMP

Self-sovereign, on-premise deployment path with no external service dependencies

CJIS

Behavioral monitoring of AI agents accessing criminal justice data systems

The Operator Desktop

See Every View of Raptor Command

Seven modules in one desktop application. Real screenshots from the shipping Electron build — no mockups, no renders. This is exactly what your security team will see on deployment day.

Shield Agents
Every connected shield with live heartbeat, source IP, ML-KEM session level, and real-time status. Click any row to drill into per-shield event history.
Threat Feed
Real-time stream of every behavioral anomaly across your deployment. Severity, attack vector, originating shield, and recommended defense action — all scored and classified.
Defense Command
One-click dispatch of ISOLATE, CONTAIN, NEUTRALIZE, and LOCKDOWN commands. Encrypted to the target shield with a sub-100-millisecond delivery target; the shield then performs the OS-level enforcement (firewall rule, process terminate, service stop).
Compliance
Six frameworks scored in real time — CMMC 2.0, HIPAA, FedRAMP, CJIS, NIST 800-171, SOC 2. Live evidence attached to every control for audit export.
Alpha Advisor
Your offline AI analyst. Reads every shield, event, and compliance control live — then answers operational questions about your specific deployment. 100% local, no data leaves the machine.
Settings
Server configuration, cryptographic fingerprints, log paths, MFA enrollment, and signature catalog management. Everything the operator needs in one screen.

Electron desktop build · Windows 10 / 11 · macOS (beta) · Linux (beta)

HYVE Alpha · Security-Domain Analyst Profiles

An AI Security Analyst, Running On Your Hardware

Alpha ships as a configured analyst profile (system prompt + sampling parameters) on top of Meta's open-weight Llama models. Four tier-scaled profiles. Model weights are Meta's upstream release under the Llama Community License. On first run, the installer pulls the right profile for your tier via Ollama (one-time internet required for the initial weight download; fully offline thereafter).

Scout Tier
HYVE Alpha-3
3 billion
parameters
Base weights: Llama 3.2 3B
Hardware: Any modern laptop / server
Footprint: 8 GB RAM · Q4_K_M quantization
Guardian Tier
HYVE Alpha-8
8 billion
parameters
Base weights: Llama 3.1 8B
Hardware: Commodity server or workstation
Footprint: 16 GB RAM · Q4_K_M quantization
Sentinel Tier
HYVE Alpha-70
70 billion
parameters
Base weights: Llama 3.3 70B
Hardware: Apple Silicon 64GB+ unified · OR · 2× GPU ≥48 GB VRAM total
Footprint: 48 GB VRAM or 64 GB unified memory
Command Tier
HYVE Alpha-405
405 billion
parameters
Base weights: Llama 3.1 405B
Hardware: 4× NVIDIA H100 80 GB · OR · equivalent multi-GPU rack
Footprint: ~230 GB VRAM total
Runtime offline
After first-run weight download, every query runs entirely on your operator server. No response or reasoning trace is ever transmitted outside the machine. Air-gapped and SCIF deployable once weights are pulled.
Security-domain prompt
Each profile ships with a configured system prompt oriented as a senior security analyst — aware of the Raptor event schema and defense command model. System prompt is HYVE-authored; model weights are Meta's upstream release.
Open-weight provenance
Built on Meta Llama 3.x under the Llama Community License. Full model provenance. No proprietary lock-in. Swap, fine-tune, or self-host alternatives at any time.
First-run setup
Installer runs "ollama pull" + "ollama create" for the profile matching your license tier. Requires one-time internet (5–60 min depending on model size). Weights cached locally thereafter.

Live context grounding (shield roster, threat feed, compliance state) is a Raptor Command Center feature — injected into each query regardless of which Alpha profile you run.
Custom domain fine-tune available as a Command-tier add-on · Contact for pricing

Zero-Touch Deployment

Buy It. Install It. Own It. Forever.

Raptor is a shipped product, not a service. You run it on your own infrastructure. The moment your license key arrives, Vibe Software Solutions is out of the loop entirely. Deployment is a standard Windows installer flow — run one EXE to stand up the server, a second to install the shield on each endpoint. Operator Manual included (15 sections, 60+ pages, every step documented).

01
Purchase online
Pick a tier. Enter org name + email. Pay with a card. Stripe-secured checkout.
02
Receive instantly
License key delivered to your inbox within seconds of payment. Installer links included.
03
Deploy in-house
Your IT team runs the installer on your server. Your license key activates it. You are the authority.
04
Own it forever
Your data, your infrastructure, your keys. Cancel any time — the software keeps running.
SYSTEM REQUIREMENTS
Operator Server
Windows 10 / 11 / Server 2016+
8 GB RAM minimum · 16 GB recommended
20 GB free disk
Client Shields
Windows 10 / 11 endpoints
4 GB RAM minimum
100 MB disk · runs as service
Network
Ports 8443 + 9443 inbound on operator
LAN or VPN to each shield
Zero external dependency
Skills Required
Familiar with Windows installers,
firewall rules, and Windows services
(standard sysadmin skill set)
📖 Read the Complete Operator Manual⬇ Download PDF
Plans

Per-Tenant. Simple Tiers.

One price per organization you protect. No per-agent fees. No per-event charges.

Scout
Up to 10 shields
$499/mo
Up to 10 raptor-shield deployments
Post-quantum threat channel (ML-KEM-768)
Raptor Command desktop app
Alpha AI Defense Advisor (local LLM)
30-day event retention
License key management
Complete operator manual (PDF)
Self-service documentation
Purchase · Deploy Yourself
MOST POPULAR
Guardian
Up to 50 shields
$1,999/mo
Up to 50 raptor-shield deployments
Post-quantum threat channel (ML-KEM-768)
Raptor Command desktop app
Alpha AI Defense Advisor (local LLM)
Multi-framework compliance (HIPAA, CMMC, FedRAMP, CJIS)
1-year event retention
Full REST API + webhook delivery
Complete operator manual (PDF)
Purchase · Deploy Yourself
Sentinel
Up to 250 shields
$7,499/mo
Up to 250 raptor-shield deployments
Raptor Command desktop app
Alpha AI Defense Advisor (local LLM)
Full 6-framework compliance engine
SIEM integration + webhook delivery
Custom retention period
Full REST API + custom rule sets
Air-gapped deployment ready
Complete operator manual (PDF)
Purchase · Deploy Yourself
GOVERNMENT
Command
Unlimited shields
$24,999/mo
Unlimited raptor-shield deployments
Raptor Command desktop app
Alpha AI Defense Advisor (local LLM)
FedRAMP / CMMC / CJIS documentation package
Air-gapped / classified / SCIF deployment ready
Source code escrow included
Perpetual license option available
Complete operator manual (PDF)
Zero post-purchase vendor involvement
Purchase · Deploy Yourself

All tiers: annual billing saves 20%. Every purchase is self-service — checkout, license key, download, deploy. No sales calls. No demos. No post-purchase vendor involvement.

Technical Overview

HYVE Raptor: Post-Quantum AI Agent Defense

An overview of the platform architecture, post-quantum transport security, AI agent behavioral monitoring, and self-sovereign licensing model.

AuthorAnthony S. Owens
OrganizationVibe Software Solutions
Versionv1.0 — 2026
Documentation

Complete Operator Manual

Every installation step, operation, defense command, compliance control, and troubleshooting scenario documented in full. Included with every license. No question left unanswered.

InstallationCommand Center ReferenceDefense CommandsAlpha AdvisorCompliance EngineFAQ + Troubleshooting
📖 Read the Manual⬇ Download PDF
Get Started

Your Agents.
Your Data.
Your Server.

Purchase, download, install, and operate — entirely on your own infrastructure. Zero vendor involvement after checkout. Your license key is generated and delivered to your inbox the moment your payment clears.

🦅 Purchase & Deploy →

Built by Vibe Software Solutions · View all HYVE products →

Made in the USA
🇺🇸 Designed & built in North Carolina
Vibe Software Solutions Inc.
Compliance & Procurement Readiness

HYVE Raptor is engineered against the control requirements of the U.S. federal regulatory frameworks most relevant to AI-agent defense deployment. The platform is designed to support organizations pursuing or maintaining compliance with:

CMMC 2.0HIPAA Security RuleFedRAMP Moderate / HighCJIS Security PolicyNIST 800-171NIST 800-53 Rev. 5SOC 2 Type IIITAR / EAR controlled

Architectural design decisions — post-quantum encryption at the application layer, zero data access for the operator, on-premises deployment with no vendor cloud dependency, DPAPI-backed cryptographic vaulting, and ML-DSA-attested shield identity — directly map to the technical safeguards required by each of the above frameworks.

Disclaimer: HYVE Raptor provides technical controls and audit evidence that support compliance obligations under the frameworks listed above. Final certification is performed by each organization's own auditor or authorizing official against its own deployment, policies, and operational procedures. Compatibility with a regulatory framework does not by itself constitute a finding of certification, accreditation, or Authority to Operate. Vibe Software Solutions Inc. makes no representation or warranty that any particular HYVE Raptor deployment will be accepted by a specific auditor, agency, or authorizing body, and is not responsible for the outcome of any external audit or certification process.