What is Milvus?

Milvus is an open-source, cloud-native vector database built for storing, indexing, and searching massive collections of vector embeddings, the numerical representations of text, images, audio, and other unstructured data that AI models produce. It's developed by Zilliz and hosted as a graduated project under the LF AI & Data Foundation.

Milvus separates compute and storage, runs as a set of independently scalable microservices on Kubernetes, and is built on top of established vector search libraries like Faiss, HNSW, DiskANN, and SCANN rather than reinventing indexing algorithms from scratch. It can run as a single-machine Standalone deployment, a fully distributed cluster for large-scale workloads, or as Milvus Lite, an embedded, pip-installable version for local development.

Typical use cases include semantic search, retrieval-augmented generation (RAG) for large language models, recommendation systems, image and video similarity search, and anomaly detection, anywhere an application needs to find the nearest neighbors of a query vector among millions or billions of stored vectors.