RepoDNA: Understand Your Codebase.

RepoDNA is an open-source tool designed to help developers understand their codebase better. It analyzes a repository, turning the discovered information into architecture maps, historical views, evidence-backed findings, reports, and more. What sets RepoDNA apart is that it does not require uploading code to a cloud service, an account, or any telemetry. Instead, it is a local-first platform that does not use AI for its analysis, making the analysis deterministic and evidence-driven.

The tool aims to provide evidence-first analysis, connecting findings to evidence such as files, lines, commits, measurements, and structures. RepoDNA distinguishes between confidence and severity, meaning it can be highly confident about something existing while it may have relatively low impact. It also keeps analysis descriptive rather than judgmental, describing the repository fingerprint without intending to grade it.

RepoDNA can analyze various aspects of a repository, including structure, languages, and architecture. In terms of structure, it identifies 25 built-in languages, providing lexical analysis, line counts, classification, source, test, generated, vendored, documentation, and configuration files. It also respects repository rules such as .gitignore and .gitattributes.

In terms of architecture, RepoDNA infers modules from package manifests and directory organization, resolves imports, builds dependency graphs, and identifies architecturally significant patterns.