This project provides an MCP (Model Context Protocol) server for interacting with code graphs, enabling advanced analysis and visualization of software repositories. It targets developers and AI assistants like ChatGPT, Cursor, and GitHub Copilot, facilitating deeper code understanding and more intelligent code-related queries.

How It Works

The server acts as a bridge between AI models and code knowledge graphs. Users can generate graphs from GitHub repositories by simply prefixing the repository URL with deepgraph.co. The MCP server then exposes various tools for querying these graphs, including semantic search for code and documentation, dependency analysis, and folder structure retrieval. This approach allows for context-aware AI interactions directly within development tools.

Quick Start & Requirements

Installation is primarily done via npx. For public graphs, the command is npx -y mcp-code-graph@latest username/repository-name. For private graphs, a CodeGPT account, API key, and optionally Organization ID and Graph ID are required, with the command structure adapting to include these credentials. Integration with tools like Gemini CLI and Claude Code involves adding specific JSON configurations or using CLI commands.

Highlighted Details

• Enables interaction with code graphs from various MCP hosts including ChatGPT, Cursor, and GitHub Copilot.
• Offers tools for semantic search of code and documentation, dependency analysis, and folder structure retrieval.
• Supports both public, no-account-required graphs and private graphs requiring a CodeGPT account.
• Provides advanced custom commands for Claude Code for comprehensive repository analysis (e.g., architecture, security, performance).

Maintenance & Community

The project is developed by JudiniLabs and associated with CodeGPT. Support and feedback can be directed to support@codegpt.co or via the app.codegpt.co website.

Licensing & Compatibility

The project is released under the MIT License, permitting commercial use and integration with closed-source projects.

Limitations & Caveats

The functionality for private graphs is dependent on having a CodeGPT account and uploading repositories to their platform. The effectiveness of advanced analysis commands relies on the quality and completeness of the generated code graph.