Deep Graph MCP Server
Enables seamless interaction with public or private code graphs from DeepGraph and CodeGPT to perform semantic searches and dependency analysis. It allows developers to explore architectural relationships, retrieve source code, and analyze repository structures across multiple AI assistants.
README
<hr>
Deep Graph MCP Server by CodeGPT
A Model Context Protocol server that enables seamless interaction with Code Graphs.
<img width="1373" alt="Screenshot 2025-06-11 at 14 01 15" src="https://github.com/user-attachments/assets/a588bafc-d5a8-4955-8d48-3addaf3ed71c" />
How this works
This MCP allows you to interact with knowledge graphs available in your CodeGPT account or with public graphs from DeepGraph.
To create a graph from any GitHub repository, simply change the URL from github.com to deepgraph.co. For example:
- GitHub repo:
https://github.com/username/repo - DeepGraph URL:
https://deepgraph.co/username/repo
<video src="https://github.com/user-attachments/assets/31f60fe6-6da4-4c67-a845-031908476351" controls autoplay></video>
You'll be able to use these advanced graph-based queries across different MCP Hosts such as ChatGPT, Cursor, Windsurf, CodeGPT Extension, GitHub Copilot, Claude Desktop, Claude Code, Gemini CLI, and others.
Available Tools
-
list-graphs: Lists available repository graphs with basic information. -
get-code: Retrieves the complete source code for a specific functionality from the graph. -
find-direct-connections: Explores the direct relationships of a functionality within the code graph. -
nodes-semantic-search: Semantically searches for code functionalities using natural language. -
docs-semantic-search: Semantically searches repository documentation. -
get-usage-dependency-links: Analyzes and lists functionalities affected by changes to a code entity. -
folder-tree-structure: Retrieves the tree structure of a folder in the repository.
For Public Code Graphs (No Account Required)
You can interact with public graphs from DeepGraph:
- Find any public repository on deepgraph.co
- Use the repository reference (e.g.,
username/repository-name)
Installation using npx
Add the following configuration to your MCP client (CodeGPT Extension, Cursor, Copilot, Claude Desktop, Windsurf, etc.):
{
"mcpServers": {
"Deep Graph MCP": {
"command": "npx",
"args": ["-y" , "mcp-code-graph@latest",
"username/repository-name", // DeepGraph repository URL
"username2/repository-name2" // Add more repositories if needed
]
}
}
}
For Private Graphs (CodeGPT Account Required)
Before using the CodeGPT MCP Server with private graphs, ensure you have:
- A CodeGPT account (sign up at app.codegpt.co)
- Uploaded a repository to Code Graph
- Get your API Key from CodeGPT API Keys page (required).
- Get your Organization ID (optional) and Graph ID (optional).
Installation with npx
Add the following configuration to your MCP client (CodeGPT Extension, Cursor, Copilot, Claude Desktop, Windsurf, etc.):
{
"mcpServers": {
"Deep Graph MCP": {
"command": "npx",
"args": ["-y" , "mcp-code-graph@latest",
"CODEGPT_API_KEY",
"CODEGPT_ORG_ID",
"CODEGPT_GRAPH_ID"
]
}
}
}
CLI Providers
Adding to Gemini CLI
Add the JSON configuration to your settings.json file:
<video src="https://github.com/user-attachments/assets/22f601dd-0564-46ac-8973-161de3017e34" controls autoplay></video>
Gemini CLI uses settings.json files for persistent configuration. There are two locations for these files:
-
User settings file: Location: ~/.gemini/settings.json (where ~ is your home directory). Scope: Applies to all Gemini CLI sessions for the current user.
-
Project settings file: Location: .gemini/settings.json within your project's root directory. Scope: Applies only when running Gemini CLI from that specific project. Project settings override user settings.
Adding to Claude Code
Follow these steps to integrate Deep Graph MCP Server with Claude Code.
Quick Setup
For Private Graphs
claude mcp add "Deep Graph MCP" npx -- -y mcp-code-graph@latest CODEGPT_API_KEY CODEGPT_ORG_ID CODEGPT_GRAPH_ID
For Public Graphs
claude mcp add "Deep Graph MCP" npx -- -y mcp-code-graph@latest username/repository-name
For team sharing, add the -s project flag:
claude mcp add -s project "Deep Graph MCP" npx -- -y mcp-code-graph@latest CODEGPT_API_KEY CODEGPT_ORG_ID CODEGPT_GRAPH_ID
# or for public graphs
claude mcp add -s project "Deep Graph MCP" npx -- -y mcp-code-graph@latest username/repository-name username2/repository-name
Verification
# Verify installation
claude mcp list
# Get server details
claude mcp get "Deep Graph MCP"
Advanced Workflows with Claude Code Custom Commands
Claude Code supports custom slash commands that combine multiple MCP Code Graph tools for comprehensive analysis workflows.
Setup
Copy the commands directory from this repository to your project root:
cp -r .claude/ /path/to/your/project/
Commit to your project's git:
git add .claude/commands/
git commit -m "Add custom Claude Code commands for Deep Graph MCP"
Available Commands
Repository-wide analysis commands (no parameters needed):
/project:analyze-architecture # Complete architectural overview
/project:security-audit # Comprehensive security analysis
/project:test-coverage-analyzer # Test coverage and quality analysis
/project:technical-debt-analyzer # Technical debt assessment
/project:api-ecosystem-analyzer # Complete API ecosystem analysis
/project:repository-onboarding # Full repository onboarding guide
Component-specific commands (require parameters):
/project:migration-planner [component/technology] # Smart migration planning
/project:performance-optimizer [component/function] # Performance optimization
/project:component-onboarding [component/feature] # Component-specific training
Usage Examples
/project:analyze-architecture
/project:migration-planner React to Vue.js
/project:performance-optimizer DatabaseService.getUserData
/project:component-onboarding authentication system
📈 Star History
Support
For support and feedback:
- Email: support@codegpt.co
- Website: app.codegpt.co
License
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。