This project provides a Model Context Protocol (MCP) plugin for AI coding assistants, enabling semantic code search across entire codebases. It targets developers using AI tools like Claude Code, Gemini CLI, Cursor, and others, offering efficient, context-aware code retrieval to enhance AI's understanding of large projects without prohibitive costs.

How It Works

The system leverages semantic search via vector databases to find relevant code snippets, moving beyond simple keyword matching. It uses Abstract Syntax Trees (AST) for intelligent code chunking and supports incremental indexing with Merkle trees for efficiency. Integration is primarily through the Model Context Protocol (MCP), allowing seamless connection with various AI coding agents.

Quick Start & Requirements

• Installation: Primarily via npx @zilliz/claude-context-mcp@latest for MCP integration.
• Prerequisites: Requires an OpenAI API key for embeddings and a Zilliz Cloud account (or Milvus instance) for vector storage.
• Setup: Involves configuring API keys and Zilliz Cloud endpoints within the AI assistant's MCP settings.

Highlighted Details

• Semantic code search for natural language queries.
• AST-based code chunking for deeper analysis.
• Incremental indexing for efficient updates.
• Supports multiple embedding providers (OpenAI, VoyageAI, Ollama, Gemini) and vector databases (Milvus, Zilliz Cloud).

Maintenance & Community

The project is maintained by Zilliz. Links to documentation, examples, and a contributing guide are provided.

Licensing & Compatibility

Licensed under the MIT License, permitting commercial use and integration with closed-source projects.

Limitations & Caveats

The project relies on external services (OpenAI, Zilliz Cloud) for core functionality, requiring API keys and potentially incurring costs. While it supports many AI assistants, compatibility depends on their MCP implementation.