Show HN: MCP Codebase Index – 87% fewer tokens when AI navigates your codebase (github.com)

The MCP Codebase Index has been introduced as a structural codebase indexing tool designed to enhance AI-assisted development. By efficiently parsing source files into metadata—such as functions, classes, imports, and dependency graphs—this tool significantly reduces the token count for AI models, achieving an impressive 87% cut. It operates with zero runtime dependencies, utilizing Python's Abstract Syntax Tree (AST) for Python files and regular expressions for TypeScript and JavaScript, making it compatible with Python 3.11 or higher. This tool is particularly significant for the AI/ML community as it streamlines code navigation for AI models like Claude Code through its 17 query tools, enabling developers to retrieve intricate project details without loading entire files. The ability to index and query codebases has profound implications for improving code understanding and productivity in AI-assisted development, potentially transforming how developers interact with large codebases. The use of persistent connections via the openclaw-mcp-adapter plugin is also notable, optimizing performance by keeping the index accessible rather than rebuilding it with each interaction.