Skip to main content
Scapius LogoScapius Architecture DSL

Enterprise Architecture as Code

Model, validate, and analyze complex systems, business capabilities, security postures, and AI agents using a modern, declarative DSL aligned with ArchiMate standards.

docker run -v $(pwd):/workspace scapiusea/scapius validate ./model.ea
architecture.ea
workspace digital_bank "Global Digital Banking" {

layer "Business Layer" {
capability cap_payments "Instant Payments" {
time "Invest"
score { business_value 10 tech_fit 9 }
}
}

layer "Application Layer" {
system sys_payment_gateway "Payment Hub" {
tier "Tier-1"
internet_facing true

app app_payment_api "Payment API" {
technology "ASP.NET Core / .NET 10"
}

agent ai_fraud_sentinel "Real-time Fraud Sentinel" {
model "gemini-2.5-flash"
autonomy_level "High"
}

relationships {
this realizes cap_payments "Realizes instant settlement"
ai_fraud_sentinel monitors app_payment_api "Inspects transactions"
}
}
}
}

Why Scapius Architecture as Code?

Treat architectural blueprints like software source code, versioned, diffable, and validated.

📐

21 ArchiMate-Aligned Entities

Map strategy, business capabilities, value streams, applications, infrastructure, and STRIDE security threats with strict Open Group ArchiMate semantic alignment.

🛡️

Security & Governance First

Native primitives for controls, policies, trust boundaries, vulnerabilities, and AI agents with real-time semantic integrity checking.

Docker & Multi-File Ready

Zero-dependency execution via the official Scapius Docker image. Seamlessly partition large enterprise estates into modular .ea files.

🎨

Expressive Style Rules

Apply custom styling, shapes, icons, and CSS classes using declarative style rules and predicate selectors without cluttering domain objects.

🧩

17 Typed Relationship Verbs

Prevent invalid topologies at compile time with strictly verified relationship verbs like composed_of, realizes, flows_to, and mitigates.

🤖

AI Agent Architecture

First-class support for cognitive AI agents, LLM foundation models, autonomy ratings, human-in-the-loop flags, and monitoring telemetry.