Overview
What is Zilliz Mcp Server?
Zilliz Mcp Server is an MCP (Model Context Protocol) server that enables AI agents to interact with Milvus, an open-source vector database, and Zilliz Cloud, its fully managed version. It allows AI assistants to create collections, insert vector data, and perform semantic searches directly within conversations, without manual database management. It is designed for developers using AI coding tools like Cursor, Claude, and Windsurf.
How to use Zilliz Mcp Server?
Configure the server in your MCP client using either Standard I/O (StdIO) via the uvx zilliz-mcp-server command with a ZILLIZ_CLOUD_TOKEN environment variable, or Streamable HTTP by cloning the repository, setting up a .env file with your Zilliz API key, and running uv run src/zilliz_mcp_server/server.py --transport streamable-http. Then your agent can leverage the exposed tools.
Key features of Zilliz Mcp Server
- Create free Milvus clusters via natural language prompts.
- Perform semantic search on vector collections.
- Manage clusters with control plane tools (list, describe, suspend, resume).
- Insert, delete, and query vector entities.
- Query real-time cluster performance metrics.
- Combine vector similarity and scalar filters in hybrid search.
Use cases of Zilliz Mcp Server
- Provision a free vector database cluster without leaving the chat interface.
- Monitor cluster health and performance conversationally.
- Search vector data semantically using plain English.
- Build and test vector search functionality within AI coding tools.
FAQ from Zilliz Mcp Server
What are the prerequisites?
Python 3.10+, uv (Python package installer), a free Zilliz Cloud account, and a Zilliz Cloud API key.
How do I obtain an API key?
Follow the Zilliz Cloud documentation on managing API keys, accessible via your Zilliz Cloud console.
What transport methods are supported?
Standard I/O (StdIO) for local agent
