中/EN
Resources
Resources

codebase-memory-mcp: A Code Intelligence Engine for AI Coding Agents, Linux Kernel Indexed in 3 Min

GitHub hot (arXiv paper) code intelligence engine. tree-sitter AST + Hybrid LSP, 158 languages, Linux kernel 28M LOC in 3 min, queries <1ms, 120× fewer tokens. Single static binary zero-dep, 15 MCP tools, 3D graph visualization. Complements graphify.

Published July 25, 20263 min read
<!-- codebase-memory-mcp-resource | resource | codebase-memory-mcp Code Intelligence Engine -->

What Is codebase-memory-mcp: A Code Intelligence Engine for AI Coding Agents

codebase-memory-mcp is a code intelligence engine for AI coding agents, full-indexing an average repo in milliseconds and the Linux kernel (28M LOC, 75K files) in 3 minutes. Structural queries in under 1ms. Single static binary for macOS / Linux / Windows. MIT license, arXiv paper (2603.27277).

Core Highlights

  • Extreme indexing speed: Linux kernel 28M LOC in 3 min. RAM-first pipeline (LZ4 compression + in-memory SQLite + Aho-Corasick pattern matching), memory released after indexing.
  • 158 languages: tree-sitter AST parsing + Hybrid LSP semantic type resolution (Python / TS / Go / C++ / Java / Rust etc., 10 langs), producing a knowledge graph of functions / classes / call chains / HTTP routes / cross-service links.
  • 120× fewer tokens: 5 structural queries ~3,400 tokens vs 412,000 via file-by-file search. One graph query replaces dozens of grep / read cycles.
  • 15 MCP tools: search / trace / architecture / impact analysis / Cypher queries / dead-code detection / cross-service HTTP linking / ADR management, etc.
  • 3D graph visualization: interactive UI at localhost:9749.
  • Zero dependencies: single static binary-download -> install -> restart agent -> done.

How to Use

One-line install (macOS / Linux):

bash
curl -fsSL https://raw.githubusercontent.com/DeusData/codebase-memory-mcp/main/install.sh | bash

Windows (PowerShell):

powershell
Invoke-WebRequest -Uri https://raw.githubusercontent.com/DeusData/codebase-memory-mcp/main/install.ps1 -OutFile install.ps1
.\install.ps1

install auto-configures detected AI agent clients (43 supported surfaces); restart the agent and use 15 MCP tools to query the code graph.

Who It's For

  • Taking over large repos: new to a project, index first and read the graph-faster than grinding code.
  • AI coding agent users: give Claude Code / Codex / Cursor code memory; agents query the graph instead of grep, saving tokens + boosting accuracy.
  • Code review / architecture: 3D graph visualization + dead-code detection + impact analysis.

vs. graphify

graphify is an AI coding skill (/graphify trigger); codebase-memory-mcp is an MCP server (resident, 15 tools). graphify is lightweight on-demand; codebase-memory-mcp is heavy full-index. They can stack.

Repo: https://github.com/DeusData/codebase-memory-mcp

Related