LLMCC is a tool that generates multi-depth architecture views of codebases to enable LLM agents to understand code structure quickly and efficiently without relying on slow grep or RAG-based methods.

It provides four levels of abstraction, from project-level relationships down to file and symbol details, allowing agents to zoom between high-level architecture and implementation specifics. The tool supports Rust and TypeScript with planned support for C++, Python, Go, and markdown, and achieves fast performance benchmarks processing large codebases in seconds.