代码知识图谱服务

CodeGraph是一款为AI助手提供深度代码理解的知识图谱服务,通过本地图数据库存储代码库信息,实现智能代码导航、影响分析和依赖跟踪。

  • by Bikach
  • Astro
  • 本地部署
  • 更新于
在 GitHub 查看 Bikach/codegraph-landing

使用说明

CodeGraph Landing Page

Landing page and documentation site for CodeGraph - an MCP server that gives AI assistants deep code understanding through knowledge graphs.

About CodeGraph

CodeGraph enhances Claude and other AI assistants with a knowledge graph of your codebase, stored in an embedded, on-disk graph database (a single local .lbug file, no server), enabling:

  • Intelligent code navigation with fewer LLM calls
  • Impact analysis on code changes
  • Dependency tracking and relationship mapping

Performance improvements:

  • 45% fewer LLM calls
  • 46% fewer tokens used
  • 41% cost reduction
  • 51% faster code analysis

Tech Stack

Technology Purpose
Astro Static site generator
Tailwind CSS Styling (GitHub Light design system)
GitHub Pages Deployment

Project Structure

src/
├── pages/
│   ├── index.astro          # Landing page
│   └── docs/                 # Documentation pages
├── components/
│   ├── landing/              # Landing page sections
│   └── docs/                 # Documentation components
├── layouts/
│   ├── Landing.astro         # Landing page layout
│   └── Docs.astro            # Documentation layout
└── styles/
    └── global.css            # Global styles and theme

Development

# Install dependencies
npm install

# Start dev server
npm run dev

# Build for production
npm run build

# Preview production build
npm run preview

Deployment

The site is automatically deployed to GitHub Pages via GitHub Actions on push to main.

Live site: https://bikach.github.io/codegraph-landing

Documentation

The site includes comprehensive documentation covering:

  • Getting Started - Overview and prerequisites
  • Installation - Claude Plugin and MCP Standalone setup
  • Commands - /codegraph:index
  • Tools - 10 graph analysis tools (search_nodes, get_callers, get_callees, get_neighbors, get_god_nodes, get_module_overview, etc.)
  • Languages - Kotlin, Java, and TypeScript/JavaScript support
  • Benchmark - Performance metrics

License

MIT