Description
Finding a function's callers in a big repo means an AI coding agent reads dozens of files into context, slow and expensive. codebase-memory-mcp indexes a codebase into a persistent knowledge graph in milliseconds for an average repo, answers structural queries in under 1ms and cuts token use by 99%.
It parses 162 languages with tree-sitter plus hybrid LSP type resolution for Python, TypeScript and more, and ships as a single binary that plugs into Claude Code, Codex, Cursor and others.
Fast indexing:The Linux kernel in 3 minutes.
Sub-ms queries:Call chains and dependencies.
162 languages:tree-sitter plus LSP semantics.
Zero dependencies:One cross-platform binary.
It parses 162 languages with tree-sitter plus hybrid LSP type resolution for Python, TypeScript and more, and ships as a single binary that plugs into Claude Code, Codex, Cursor and others.
Features
Fast indexing:The Linux kernel in 3 minutes.
Sub-ms queries:Call chains and dependencies.
162 languages:tree-sitter plus LSP semantics.
Zero dependencies:One cross-platform binary.
