Sruja is a next-generation architecture-as-code language with first-class support for requirements, ADRs, and extensible validation.

⚠️ Alpha Release (v0.1.0): Sruja is under active development. APIs may change. See ROADMAP.md for the path to v1.0.0.

https://sruja.ai

• 🎯 Architecture DSL: Define systems, containers, components, and relations
• ✅ Validation Engine: Cycle detection, orphan detection, unique IDs, valid references
• 📊 D2 Export: Export to D2 diagrams for rendering
• 🎨 Code Formatter: Auto-format your architecture with sruja fmt
• 🌳 Tree View: Visualize hierarchy with sruja tree

sruja/
├── cmd/
│   └── sruja/            # Main CLI tool
├── pkg/                  # Shared Go packages
│   ├── engine/           # Validation engine
│   ├── language/         # Parser, AST, lexer
│   └── export/           # Exporters (D2, etc.)
└── examples/             # Example .sruja files

curl -fsSL https://raw.githubusercontent.com/sruja-ai/sruja/main/scripts/install.sh | bash

Download the latest release from GitHub Releases.

go install github.com/sruja-ai/sruja/cmd/sruja@latest

Create example.sruja:

architecture "My System" {
    system App "My App" {
        container Web "Web Server"
        datastore DB "Database"
    }
    person User "User"

    User -> Web "Visits"
    Web -> DB "Reads/Writes"
}

Export to D2:

sruja export d2 example.sruja

Lint your code:

sruja lint example.sruja

Format your code:

sruja fmt example.sruja

View hierarchy:

sruja tree --file example.sruja

• Go 1.25+

# Install dependencies
go mod download

# Setup git hooks (recommended)
make setup-hooks

# Build CLI
make build

A pre-commit hook automatically tests code compilation when you commit changes to the learn/ directory. This prevents broken code from being committed.

Setup:

make setup-hooks

The hook will:

• Test playground examples compile correctly
• Test course code blocks compile correctly
• Test docs code blocks compile correctly
• Block commits if any code fails to compile

See docs/DEVELOPMENT.md for more details.

We welcome contributions of all sizes — from small doc fixes to core features.

• Start with issues and "good first issues": good first issue Good for newcomers
• Read the contribution guide

• Run local checks: make test, make fmt, make lint
• Add/update tests for new behavior
• Keep changes focused and well‑scoped
• Use Conventional Commits (e.g., feat: …, fix: …, docs: …)

MIT