aut0arch Documentation

Welcome to the official documentation for aut0arch!

aut0arch is an automated architecture discovery and explanation pipeline. It is designed to take a repository of source code (currently focusing on Java) and automatically reverse-engineer its high-level architecture. It doesn’t just draw diagrams; it leverages Large Language Models (LLMs) to intelligently explain what each component does, why it exists, and how it interacts with the rest of the system.

Why aut0arch?

Understanding a new or legacy codebase is one of the most time-consuming tasks for software engineers. Traditional documentation is quickly outdated, and raw dependency graphs are often too granular to provide meaningful insight.

aut0arch solves this by:

1. Parsing the actual source code to build a ground-truth semantic graph.
2. Clustering those fine-grained details into high-level, cohesive modules.
3. Explaining those modules using AI, giving them human-readable roles and context.
4. Visualizing the result in an interactive web application.

Repository Structure

The aut0arch ecosystem consists of 5 specialized microservices:

• server: The central orchestrator that manages the pipeline.
• parser: The engine that reads code and builds the raw dependency graph.
• clustering: The tool that abstracts the raw graph into architectural components.
• explainer: The AI-integration layer that writes human-readable descriptions.
• app: The React frontend that visualizes the final architecture.

To get started, please refer to the Installation via Docker guide.

To learn more about how they work together, head over to the Architecture Overview.