mcp-server-milvus
This project provides a Model Context Protocol (MCP) server that integrates with the Milvus vector database, offering tools and features for AI-powered applications such as Claude Desktop and Cursor. It supports efficient data operations, collection management, and sophisticated search capabilities.
MCP Server for Milvus
The Model Context Protocol (MCP) is an open protocol that enables seamless integration between LLM applications and external data sources and tools. Whether you're building an AI-powered IDE or creating custom AI workflows, MCP provides a standardized way to connect LLMs with the context they need.
This repository contains a MCP server that provides access to Milvus vector database functionality.
Prerequisites
- Python 3.10 or higher
- A running Milvus instance (local or remote)
- uv installed (recommended for running the server)
Usage
The recommended way to use this MCP server is to run it directly with uv without installation.
Supported Applications
- Claude Desktop
- Cursor
- Custom MCP clients
Available Tools
Search and Query Operations
- Search and query documents using vector or full-text search
- Perform hybrid searches combining vector similarity and attribute filtering
Collection Management
- Manage collections: list, create, load, release, and get statistics
Data Operations
- Insert, upsert, and delete data
- Manage dynamic fields
Index Management
- Create and manage indexes on vector fields
Environment Variables
Configure Milvus server URI, token, and database name.
Troubleshooting
Common Issues
- Connection and authentication errors
- Tool not found issues in Claude Desktop or Cursor
Getting Help
Join the Zilliz Community Discord for support or check GitHub Issues.