MCP Server for Milvus

This repository hosts an MCP (Model Context Protocol) server, facilitating seamless integration between LLM applications and Milvus, a vector database. MCP standardizes how LLMs connect with external data sources, enhancing AI-powered IDEs, chat interfaces, and custom AI workflows.

The server, designed to run with uv, requires Python 3.10+ and a running Milvus instance. It supports applications like Claude Desktop and Cursor, enabling them to leverage Milvus's capabilities. Configuration involves specifying the server path and Milvus URI in the application settings.

Key features include vector and text search, collection management (create, load, release), and data operations (insert, delete). Available tools include milvus_text_search, milvus_vector_search, and milvus_create_collection. Environment variables like MILVUS_URI and MILVUS_TOKEN allow for flexible configuration. Troubleshooting tips address common connection and authentication issues.