CodeGraph MCP
CodeGraph MCP is a powerful standalone tool that parses your entire C/C++ codebase into a semantic knowledge graph and seamlessly exposes it to AI coding assistants via the Model Context Protocol (MCP). By providing AI (like Claude Desktop, Cursor, or Google Antigravity) with a structural map of your project—including caller/callee relationships, file dependencies, and dynamic function definition
README
<div align="center"> <h1>CodeGraph MCP</h1> <p><em>The Ultimate Codebase Context Engine for AI Assistants (Cursor, Claude, Antigravity)</em></p>
🚀 Overview
CodeGraph MCP is a powerful standalone tool that parses your entire C/C++ codebase into a semantic knowledge graph and seamlessly exposes it to AI coding assistants via the Model Context Protocol (MCP).
By providing AI (like Claude Desktop, Cursor, or Google Antigravity) with a structural map of your project—including caller/callee relationships, file dependencies, and dynamic function definitions—CodeGraph dramatically reduces AI hallucinations and improves the accuracy of codebase-wide refactoring.
⚡ Breakthrough Performance (v1.0.0 Commercial)
The new architecture replaces the legacy Python AST with optimized C++ binaries, unlocking unprecedented power:
| Benchmark Metric | Real-world Result | The Difference |
|---|---|---|
| Scan Speed (760+ files) | < 1.2 seconds | 10x faster than legacy AST. Near-zero latency parsing. |
| Tracing Accuracy | 100% Deterministic | No probabilistic guessing (unlike Vector search). 100% accurate Call & Inherit tracking. |
| RAM Consumption | ~12 MB RAM | Extremely lightweight. No background Database or Vector Store required. |
| Source Code Security | 100% Offline | Your code never leaves your machine. The tool itself is protected by PyArmor anti-tampering. |
🌟 Advantages
- 🎯 Impact Analysis: Before the AI edits Function A, CodeGraph traces 3 levels deep to warn the AI: "Watch out! Changing this will break Class B in file X and Test C in file Y".
- 📦 Context Packing: Automatically calculates dependencies around your editing point and packs them perfectly into the Token limit. Stop dumping garbage into your prompt.
- 🕸️ Obsidian-style Graph View: Renders your architecture into a highly interactive HTML5 canvas. Observe thousands of file connections with just a scroll of your mouse.
- 🔌 Plug-and-Play (MCP): Standardized as a native MCP Server. Works instantly out-of-the-box with Cursor, Windsurf, or Claude Desktop.
📥 Download & Installation
You can download the latest official release directly from GitHub:
👉 Download CodeGraph v1.0.0 (30-Day Free Trial)
📖 Setup Instructions
- Extract the ZIP file to a permanent location (e.g.,
C:\CodeGraph). - Double-click
CodeGraph.exeto start your 30-day trial and generate a structural map for your project. - For detailed MCP integration instructions with Cursor, Claude, and Antigravity, please read the bundled
USER_GUIDE.mdinside the ZIP file.
🛒 Licensing
CodeGraph comes with a fully functional 30-day free trial.
After the trial expires, you can purchase a permanent license key to continue using the software: 👉 Buy License Key on Lemon Squeezy ($15)
(One license supports up to 2 devices simultaneously).
🌐 Website
Visit our official landing page for more information and interactive demos: 👉 CodeGraph Official Website
© 2026 CodeGraph. All rights reserved.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。