Skip to content

Architecture

The Pluggable Brain

Breeze exposes its RMM capabilities as a standardized tool catalog. Any AI brain — BYOK or LanternOps — connects through the same interface. Your RMM stays sovereign. You're never locked in.

Deploy Now View Pricing

Architecture

How It Works 

┌─── LanternOps Cloud (or BYOK Local) ───┐
│  Claude Agent SDK                        │
│  Decides what to investigate and act on  │
│         │                                │
│    Calls tools ──→ HTTP or Local         │
└─────────┼────────────────────────────────┘
          │
┌─────────▼────────────────────────────────┐
│  Breeze RMM — Brain Connector API        │
│  ┌──────────────────────────────────┐    │
│  │ Risk Engine (ALWAYS enforced)    │    │
│  └──────────────────────────────────┘    │
│  Tool Catalog → Devices, Alerts,         │
│  Scripts, Patches, Reports               │
└──────────────────────────────────────────┘

Tool Catalog IS the API

The RMM registers its capabilities as tool definitions. Both BYOK and LanternOps consume the same schemas. Change a tool — that's a versioned API change.

Risk Engine Enforces Locally

The RMM classifies every action into 4 risk tiers and enforces approval requirements. The brain can request anything — the RMM decides whether to allow it.

Events Trigger, Tools Act

The RMM pushes structured events (alerts, threshold breaches, security events). The brain responds by calling tools. Clean separation of concerns.

Modes

Two Modes, One Platform

BYOK Mode

Free

Included with every Breeze deployment

  • Your Anthropic API key
  • Single-agent reactive loop
  • Handles one event at a time
  • Tools call local functions directly
  • Basic risk classification
  • No persistent memory

Good enough to be useful. Limited enough to upgrade.

LanternOps Brain

Managed

The commercial brain — everything BYOK can't do

  • Multi-agent orchestration (triage → remediation → compliance)
  • Persistent memory per tenant
  • Cross-tenant intelligence (anonymized patterns across all managed tenants)
  • Automated playbook library
  • Compliance evidence generation (NIST, CIS, SOC 2, HIPAA)
  • Proactive anomaly detection

Same tools. Smarter brain. Zero migration.

Upgrade Path

The Upgrade Path

1

Deploy Breeze

Free, open-source. Full RMM with 44 modules.

2

Enable BYOK AI

Add your Anthropic API key. AI triage starts immediately.

3

Connect LanternOps

One click in Settings → AI Brain → LanternOps. OAuth flow.

4

Scale with Intelligence

Persistent memory, cross-tenant patterns, automated playbooks.

Trust Model

Trust Model

Your RMM is always the final authority

LanternOps can request actions. Your Breeze instance decides whether to allow them. Risk classification is enforced at the RMM level — not the brain level. Double validation: LanternOps pre-validates with its intelligence layer, then your RMM re-validates with its risk engine. Belt and suspenders.

Every action is auditable. Every tool call is logged. Every approval decision is recorded.

Ready to get started?

Deploy Breeze for free, enable AI triage with your own key, and upgrade to LanternOps when you're ready for multi-agent intelligence.

Deploy Breeze View Pricing Learn about LanternOps