Milvus MCP

Name: Milvus MCP
Author: Zilliz

by Zilliz

Semantic search and vector operations on Milvus vector database from AI assistants

database Python Intermediate Self-hostable No API key Verified

⭐ 200 stars 📅 Updated: 1mo ago

View on GitHub View package

Description

MCP server for interacting with Milvus, the open-source vector database designed for AI applications. Enables LLMs to perform semantic search, manage collections, insert and query vectors, and explore embedding data through natural language. Supports both self-hosted Milvus instances and Zilliz Cloud managed service. Build RAG pipelines, explore similarity relationships, and manage your vector data — all through your AI assistant. Ideal for teams building AI-powered search, recommendation systems, and knowledge bases that need vector similarity matching at scale.

✅ Best for

Teams building RAG, semantic search, or recommendation systems with Milvus

⏭️ Skip if

You need a traditional relational database connector or full-text search only

💡 Use cases

Building and querying RAG pipelines with semantic vector search
Managing vector collections and inserting embeddings from AI workflows
Exploring similarity relationships between documents or data points
Prototyping recommendation systems with vector-based matching

👍 Pros

✓ Supports both self-hosted Milvus and Zilliz Cloud for flexibility
✓ Full collection lifecycle management — create, describe, list, and delete
✓ Semantic search with filtering and configurable similarity metrics
✓ No API key required for self-hosted Milvus — quick local setup

👎 Cons

✗ Requires a running Milvus instance or Zilliz Cloud account
✗ Vector operations require pre-computed embeddings — no built-in embedding generation
✗ Complex vector queries with many dimensions may produce verbose results

🔧 Exposed tools (5 tools)

Tool	Category	Description
insert_vectors	data	Insert vector embeddings into a collection
list_collections	discovery	List all available collections in the Milvus instance
describe_collection	discovery	Get the schema and details of a specific collection
create_collection	management	Create a new vector collection with a specified schema
search_vectors	query	Perform semantic similarity search on a vector collection

⚡ Installation

Prerequisites:

• python v3.10+
• Milvus instance (self-hosted or Zilliz Cloud, optional API key for cloud)

Check Claude Code documentation to configure this MCP server.

💡 Tips & tricks

For local development, start Milvus with Docker using the official docker-compose. Set MILVUS_URI to point to your instance. For Zilliz Cloud, configure ZILLIZ_CLOUD_URI and ZILLIZ_CLOUD_TOKEN. Use create_collection to set up your schema before inserting vectors.

🔄 Alternatives

Chroma MCP

Vector search, full-text search, and metadata filtering for AI via Chroma's embedding database

Similar functionality

Qdrant MCP

Semantic memory layer for AI agents with vector search via Qdrant