milvus
Milvus is a high-performance, cloud-native vector database built for scalable vector ANN search
What is milvus?
Milvus is a high-performance, cloud-native vector database built for scalable vector ANN search
Overview of milvus
What it is
Milvus is a vector database built for scale, written in Go and C++. It organizes and searches large amounts of unstructured data such as text, images, and multi-modal information. It uses a fully distributed, Kubernetes-native architecture and also supports Standalone mode for single-machine deployment. Milvus Lite is a lightweight version for quickstart in Python via pip install. The project is under the LF AI & Data Foundation, with Zilliz as its major contributor. It stores vectors together with scalar data types such as integers, strings, and JSON objects, and supports vector search with metadata filtering or hybrid search.
Key features
- Distributed architecture that separates compute and storage, with independent scaling of query nodes for read-heavy workloads and data nodes for write-heavy workloads; stateless microservices on Kubernetes and replica support for fault tolerance and throughput.
- Support for major vector index types including HNSW, IVF, FLAT (brute-force), SCANN, and DiskANN, with quantization-based variations and mmap.
- Hardware acceleration for CPU/GPU, including GPU indexing such as NVIDIA's CAGRA.
- Sparse vectors for full text search with BM25 and learned sparse embeddings such as SPLADE and BGE-M3, with functions to rerank results from multiple search requests.
- Multi-tenancy through isolation at database, collection, partition, or partition key level.
- Hot/cold storage, keeping frequently accessed data in memory or on SSDs and less-accessed data on slower, cost-effective storage.
- Mandatory user authentication, TLS encryption, and Role-Based Access Control (RBAC).
Use cases
The README's tutorials cover Retrieval-Augmented Generation (RAG), semantic search, full text search, hybrid search, question answering, recommendation systems, image search, multimodal search with multi vectors, graph RAG, vector visualization, and clustering. Milvus is described as being used for text and image search, RAG, and recommendation systems.
Getting started
python pip install -U pymilvus python from pymilvus import MilvusClient # Local vector database with Milvus Lite (pip install pymilvus[milvus-lite]) client = MilvusClient(uri, token) client.create_collection( collection_name, dimension=768, ) res = client.insert(collection_name, data) res = client.search( collection_name, data=query_vectors, limit=2, output_fields=[], )
License and pricing
Milvus is open source under the Apache 2.0 License and is distributed under the LF AI & Data Foundation, with Zilliz as its major contributor. The README states that Milvus is available as a fully managed service on Zilliz Cloud, with Serverless, Dedicated, and BYOC options; no pricing is stated in the README or metadata.
Based on the project's public README and repository metadata.