Milvus

Milvus is a high-performance vector database built for scale. It stores and searches unstructured data such as text, images, and multimodal information by similarity, powering AI applications like semantic search, retrieval-augmented generation, and recommendation systems.

Written in Go and C++, Milvus uses a fully distributed, Kubernetes-native architecture that separates compute and storage. It scales horizontally to billions of vectors, keeps data fresh with real-time streaming updates, and accelerates search on CPU and GPU. It supports index types including HNSW, IVF, FLAT, SCANN, and DiskANN, plus metadata filtering and hybrid search combining dense vectors with BM25 full-text search.

Milvus is an LF AI & Data Foundation project under the Apache 2.0 license, with Zilliz as its major contributor. It runs standalone on a single machine, embeds via Milvus Lite for Python, or runs fully managed on Zilliz Cloud.