Open-source cloud-native vector database built for billion-scale similarity search
Milvus documents a high-performance vector database at milvus.io/docs for storing, indexing, and searching embedding vectors with metadata filtering and hybrid search. Deployment options include Milvus Lite (`pip install pymilvus` for notebooks/edge), Milvus Standalone (single Docker image), and Milvus Distributed on Kubernetes per milvus.io/docs/v2.6.x/install-overview. Official SDKs include PyMilvus, Go, Java, Node.js, and C#; Zilliz Cloud offers managed Milvus. Architecture separates access, coordinator, worker, and storage layers with object storage backends (MinIO, S3, Azure Blob) per milvus.io/docs/architecture_overview.
Use cases
- Production RAG catalogs at billion-vector scale on Kubernetes
- Recommendation systems combining vector similarity with structured filters
- Notebook prototyping with Milvus Lite then migrating to Standalone/Distributed
- Agent memory layers paired with zilliztech/mcp-server-milvus
- Multimodal embedding search when combined with external embedders
Key features
- HNSW, DiskANN, and other ANN indexes with scalar/JSON metadata filtering
- Milvus Lite, Standalone, and Distributed deployment modes
- Hybrid dense-sparse search and multi-vector support in recent releases
- PyMilvus MilvusClient API for collections, insert, search, and query
- LF AI & Data Foundation project with Zilliz as core maintainer
Who Is It For?
- ML engineers operating large-scale vector search infrastructure
- Platform teams evaluating open-source alternatives to single-vendor vector clouds
- Developers prototyping locally with Milvus Lite before production rollout
Frequently Asked Questions
- Is Milvus the same as Zilliz Cloud?
- Milvus is the open-source project; Zilliz Cloud is the fully managed service built on Milvus.
- Which Python client should I use?
- Docs recommend PyMilvus with MilvusClient for current releases—see milvus.io/docs and pymilvus docs.
- How do agents connect?
- Zilliz maintains mcp-server-milvus (documented at milvus.io/docs/milvus_and_mcp) for MCP clients.
Related
Related
3 Indexed items
Weaviate
Weaviate documents an open-source vector database at docs.weaviate.io/weaviate for storing objects and vector embeddings with semantic, keyword, and hybrid search, RAG, reranking, and agent workflows. The ecosystem includes self-hosted Docker/Kubernetes installs, Weaviate Cloud (console.weaviate.cloud), Query Agent, and Weaviate Embeddings for managed inference. Client libraries include Python (`weaviate-client` v4, requires Weaviate 1.23.7+), TypeScript, Go, and Java with REST, gRPC, and GraphQL APIs per the official documentation.
Pinecone
Pinecone documents a fully managed vector database at docs.pinecone.io for storing, indexing, and querying high-dimensional embeddings at production scale. Serverless indexes support document schemas mixing dense vectors, sparse vectors, and full-text search fields with metadata filtering per docs.pinecone.io/guides/get-started/concepts. Official SDKs include Python, Node.js, Java, and Go; REST API access uses documented rate limits and plan tiers (Starter, Standard, Enterprise). Pinecone also documents Pinecone Assistant, Dedicated Read Nodes, BYOC, and Nexus offerings on pinecone.io alongside MCP integrations (Pinecone MCP Server and Pinecone Docs MCP Server) for agent workflows.
Qdrant
Qdrant documents an AI-native vector search engine at qdrant.tech/documentation for storing, indexing, and querying high-dimensional vectors with optional payloads, supporting dense, sparse, and multi-vector configurations. Official guides cover Docker/Kubernetes self-hosting, Qdrant Cloud on AWS/GCP/Azure, Hybrid Cloud, Private Cloud, and Qdrant Edge for embedded retrieval. Client libraries include Python (`qdrant-client`), JavaScript/TypeScript (`@qdrant/js-client-rest`), Rust, Go, Java, and .NET with REST and gRPC APIs per the API reference at api.qdrant.tech.