Milvus

Milvus is a widely adopted open-source vector database built for billion-scale similarity search, with a distributed, cloud-native architecture that separates compute and storage. It is free to self-host under Apache 2.0, with a managed option via Zilliz Cloud. It targets teams that need production-grade vector search at large scale.

Milvus is one of the most widely adopted open-source vector databases, built for high-performance similarity search at very large scale. Created by Zilliz and now a graduated project of the LF AI & Data Foundation, it features a distributed, cloud-native architecture that separates compute and storage, enabling it to handle billions of vectors across horizontally scaled clusters. It supports multiple index types, GPU acceleration, hybrid search, and rich filtering. Milvus targets teams that need production-grade vector search at scale and are willing to run real infrastructure to get it. Its architecture (with components for querying, data, indexing, and coordination) is powerful but more involved than lightweight embedded databases, which is the tradeoff for its scalability and flexibility. Milvus Lite offers a lightweight embedded mode for smaller local use and prototyping. The project is free and open source under Apache 2.0. For teams that prefer not to operate the distributed system themselves, Zilliz offers Zilliz Cloud, a fully managed service (serverless, dedicated, and enterprise tiers) billed by compute units and storage. Milvus competes with Pinecone, Weaviate, Qdrant, and Chroma, differentiating on proven scale and a mature, cloud-native design.