Milvus
Zilliz
Star46.5KBook a DemoTry Managed Milvus

Milvus solutions

Choose a Milvus architecture and retrieval approach for your workload, with examples, documentation and customer stories.

Database for AI agents

Milvus for AI agents: scoped memory, RBAC, hybrid retrieval and regex filters. Zilliz Cloud enterprise edition adds row-level security.

Distributed architecture

Milvus architecture explained: storage-compute separation, stateless nodes on object storage, the write and search paths, plus RaBitQ, DiskANN and tiered storage.

Hybrid search

Hybrid search in one open-source vector database: built-in BM25 full-text search, dense and sparse vectors, RRF or weighted fusion, and filters in one query.

Open-source vector database

Milvus is a free, Apache 2.0 open source vector database with 40K+ GitHub stars. Compare it with Qdrant, Weaviate, Chroma and pgvector; run via pip, Docker or K8s.

Search on the lake

Vector search on data lake tables without copying data: Milvus external collections index Lance datasets and Apache Iceberg tables in place and serve them.

Vector database for RAG

Milvus is an open-source vector database for RAG: BM25 + dense hybrid search, metadata filtering, reranking, multi-tenancy, disk indexes, Lite to billions of chunks.