Skip to content

Extractors

Extractors are the eyes of Archlette. They analyze source code, configuration files, and infrastructure definitions to extract architectural information.

Each extractor understands a specific language or framework. Pick the ones that match your stack.

Available Extractors

Basic Node

Path: extractors/builtin/basic-node

Analyzes TypeScript and JavaScript codebases using JSDoc annotations.

Extracts:

  • Components (modules, namespaces)
  • Actors (users, external systems)
  • Relationships (dependencies between components)
  • Code elements (classes, functions)

Best for: Application code, services, APIs, libraries

→ Full documentation

Basic Wrangler

Path: extractors/builtin/basic-wrangler

Analyzes Cloudflare Wrangler configuration files to extract deployment topology.

Extracts:

  • Containers (Cloudflare Workers)
  • Deployments (production, staging, preview environments)
  • Service bindings (worker-to-worker dependencies)
  • Infrastructure bindings (KV, R2, D1, Durable Objects, Queues)

Best for: Cloudflare Workers microservices architecture

→ Full documentation

Comparison

Feature basic-node basic-wrangler
Language TypeScript/JavaScript TOML
Source Code annotations Config files
Extracts Components ✅ Yes ❌ No
Extracts Containers ⚠️ Optional ✅ Yes
Extracts Deployments ❌ No ✅ Yes
Extracts Relationships ✅ Yes ✅ Yes
Infrastructure Aware ❌ No ✅ Yes

Configuration Pattern

Typical setup:

extractors:
  - use: extractors/builtin/basic-node
    inputs:
      include: ['src/**/*.ts']
      exclude: ['**/*.test.ts']

All extractors support:

  • use — Path to extractor module
  • name — Optional friendly name
  • inputs — File patterns (include/exclude)
  • props — Extractor-specific configuration

Multiple Extractors

Extractors run in sequence. Outputs merge into a single Intermediate Representation (IR).

Example: Full-stack application

extractors:
  # Application code
  - use: extractors/builtin/basic-node
    name: app-code
    inputs:
      include: ['src/**/*.ts']

  # Deployment topology
  - use: extractors/builtin/basic-wrangler
    name: worker-deployment
    inputs:
      include: ['iac/workers/**/*.toml']

Result: One architecture with both logical components (from code) and physical deployments (from config).

Coming Soon

Planned extractors:

  • 🔸 basic-python — Python (Django, Flask, FastAPI)
  • 🔸 basic-terraform — Infrastructure as Code (AWS, Azure, GCP)
  • 🔸 basic-docker — Docker Compose
  • 🔸 basic-openapi — REST API specifications
  • 🔸 basic-graphql — GraphQL schemas

Build Your Own

Creating custom extractors is straightforward:

  1. Implement the ArchletteExtractor interface
  2. Return ArchletteIR conforming to the schema
  3. Register in your .aac.yaml

→ Plugin Development Guide

Need Help?

Choosing an extractor? → Decision guide
Configuration options? → Configuration reference
Want to contribute? → Contributing guide