@AriXZone: The Linux kernel, a very large project with 28 million lines of code and 75,000 files, takes only about 3 minutes to index, with sub-millisecond query responses. codebase-memory-mcp is an open-source MCP server from DeusData that provides AI coding assistants (Claude Cod…

The Linux kernel, a very large project with 28 million lines of code and 75,000 files, takes only about 3 minutes to index, with sub-millisecond query responses. codebase-memory-mcp is an open-source MCP server from DeusData that provides AI coding assistants (Claude Code, Cursor, and 11 other agents) with code structure understanding. It pre-parses code repositories into persistent knowledge graphs, allowing agents to use graph queries instead of per-file reading and grep. Key Features: Based on tree-sitter, it supports parsing 158 languages, with mainstream languages (Python/TS/JS/Go/Java/C++/Rust, etc.) additionally using Hybrid LSP for semantic type resolution, building graphs of functions, classes, call chains, HTTP routes, and cross-service relationships. Provides 14 MCP tools, including code search, call chain tracing, architecture overview, impact analysis, Cypher query, dead code detection, semantic search, etc. Highly efficient: for the same structured query tasks, token consumption is reduced by about 99% compared to traditional grep; indexing is fast, completing in seconds to minutes for typical repositories. Single static binary, zero runtime dependencies, no API Key required, no built-in LLM—only graph construction and querying; inference is still done by the calling agent. All data is processed locally, with no uploads or telemetry. Supports team sharing of compressed graph snapshots to reduce redundant indexing; provides an optional 3D graph visualization UI. Security includes SLSA L3 build provenance, Sigstore signing, SHA-256 checksums, and CodeQL scanning. Installation: one-click script on macOS/Linux, or let Claude Code directly pull the repository and build automatically; Windows requires WSL2.