Code Index MCP

What is Code Index MCP?

Code Index MCP is a Model Context Protocol server that provides intelligent code indexing, advanced search, and detailed code analysis for AI assistants. It bridges the gap between AI models and complex codebases, supporting code review, refactoring, documentation generation, debugging, and architectural analysis.

How to use Code Index MCP?

Install via uvx (recommended) or pip install code-index-mcp. Add the server to your MCP configuration (e.g., claude_desktop_config.json) with the command uvx code-index-mcp and optionally the --project-path argument to auto-initialize a repository. After restarting the application, use tools like set_project_path, search_code_advanced, find_files, and get_file_summary to interact with your codebase.

Key features of Code Index MCP

• Dual-strategy parsing: tree-sitter AST for 10 core languages, fallback for 50+ file types
• Advanced search with auto-detection of ugrep, ripgrep, ag, or grep
• Real-time file monitoring with automatic index updates
• Persistent caching for fast subsequent access
• Multi-language support including Python, JavaScript, TypeScript, Java, Kotlin, C#, Go, Rust, Objective-C, Zig

Use cases of Code Index MCP

• Code review: find all places using an old API
• Refactoring: locate where a function is called across the project
• Learning: explore main components of a React project
• Debugging: search for all error handling related code

FAQ from Code Index MCP

What Python version is required?

Python 3.10 or higher is required.

How do I set the project path for indexing?

You can either pass --project-path /absolute/path/to/repo to the args array in the MCP configuration, or use the set_project_path tool after startup.

Which languages get deep tree-sitter AST parsing?

Ten languages: Python, JavaScript, TypeScript, Java, Kotlin, C#, Go, Rust, Objective-C, and Zig. All other programming languages use a fallback strategy for basic file indexing.

Does Code Index MCP support real-time file changes?

Yes, it includes a file watcher that automatically updates the index when files change. You can check status with get_file_watcher_status and configure it with configure_file_watcher.

How does the search tool work?

search_code_advanced uses literal matching by default, with optional regex=True, fuzzy matching, file filtering, and paginated results. Regex mode requires a native search tool (ugrep, ripgrep, ag, or grep) to be installed.