Show HN: Carto – structural intelligence for AI coding agents (OSS) (github.com)

Carto, a newly released open-source tool, introduces a structural intelligence layer for AI coding agents, significantly enhancing their understanding of codebases. Unlike conventional coding tools that treat files as text, Carto maps out the architecture of a project by identifying domains, import graphs, and related routes. This allows AI coding agents to generate precise change plans rather than making educated guesses, leading to more efficient code modifications and better overall integration with various AI development environments such as VS Code, JetBrains, and Claude. The significance of Carto lies in its ability to transform how AI tools interact with code by indexing structural elements and dependencies, yielding instant feedback on implications of code changes. For example, when a user queries to "add rate limiting to /api/users," Carto provides a comprehensive overview, detailing affected routes, related files, and potential impact, rather than simply searching for text matches. By leveraging advanced techniques like Leiden+CPM graph clustering and SQLite for data management, it effectively supports various programming languages and frameworks, ensuring quick and accurate indexing. This innovation positions Carto as a vital resource for developers and AI practitioners, streamlining code understanding and enhancement in complex projects.