Code-mapper: Free CLI tool to reduce LLM token usage on any codebases (github.com)

A new open-source command-line tool called Code-mapper has been released, designed to significantly reduce token usage when interacting with large codebases in language models (LLMs). By creating a compact `PROJECT_CONTEXT.md` file, Code-mapper allows an LLM to understand a codebase's architecture with an approximate reduction in token usage of 78%. For instance, in a 4,200-line project, the tool condenses around 21,050 tokens typically used from raw source files down to just 4,711 tokens by summarizing classes, functions, and dependencies. This innovation is significant for the AI/ML community as it utilizes Mermaid, a highly efficient visual format, to present complex code structures in a token-dense manner, accelerating LLM comprehension and interaction speed. Code-mapper supports Python natively, applying built-in tools for accurate parsing, while also offering compatibility with other languages via regex parsing. Future enhancements could include better parsing for TypeScript and other languages, improved handling of large codebases, and integration with IDEs for real-time updates. Overall, this tool streamlines how LLMs consume code and sets a precedent for managing token costs in AI-assisted software development.