Categories
Open Source AI Machine Learning Cloud Native Go Kubernetes Vector Database Similarity SearchStars
Forks
Watchers
Developer links
Milvus
With over 45,000 GitHub stars and 100 million Docker pulls, Milvus is the most widely adopted open-source vector database, powering production AI systems at NVIDIA, Salesforce, eBay, Airbnb, and DoorDash. The distributed architecture separates compute and storage with stateless microservices on Kubernetes, horizontally scaling query nodes for read-heavy workloads and data nodes for write-heavy ingestion independently. Milvus 3.0 introduces lake-native retrieval that builds and serves indexes directly over vector data in object storage and open formats including Parquet, Lance, Iceberg, and Vortex without maintaining separate copies. Native hybrid search unifies lexical BM25 full-text retrieval and semantic vector search in a single engine with metadata filtering, eliminating the need for separate search infrastructure. Hardware-accelerated ANN indexing supports IVF, HNSW, DiskANN, and GPU-based indexes with BitQ 1-bit quantization cutting memory usage by 72 percent. SDKs for Python, Go, Node.js, and Java provide programmatic access, while Milvus Lite offers lightweight embedding for local development via pip install. Server-side aggregation, sorting, faceted search, StructArray for nested document structures, and ColBERT multi-vector scoring move ranking and result processing into the engine. The Path Index enables 100x faster JSON filtering with support for 100,000+ collections per cluster for multi-tenant deployments. Self-hosting deploys via Docker Standalone or Kubernetes with Helm charts using S3-compatible, GCS, or Azure Blob storage backends. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache 2.0 licensed.
Benefits
- Billion-Scale Vector Search Performance
- Hardware-accelerated ANN indexing with IVF, HNSW, DiskANN, and GPU-based indexes delivers sub-10ms latency on billion-scale vector collections with horizontal scaling on Kubernetes.
- Native Hybrid Search Engine
- Unified BM25 full-text retrieval and semantic vector search with metadata filtering in a single engine eliminates maintaining separate search infrastructure for keyword and semantic queries.
- Lake-Native Vector Retrieval
- Milvus 3.0 builds and serves indexes over vector data in object storage and open formats like Parquet, Lance, and Iceberg without copying data into a separate database.
- Decoupled Compute and Storage
- Stateless microservices architecture on Kubernetes independently scales query and data nodes, providing quick recovery from failures and optimal resource utilization per workload type.
Features
- Multi-Vector ColBERT Scoring
- StructArray support for nested document-chunk structures and ColBERT multi-vector scoring move complex ranking and reranking operations directly into the engine.
- BitQ Memory Optimization
- 1-bit quantization reduces memory usage by 72 percent while maintaining search quality, enabling cost-effective deployment of large-scale vector collections in production.
- 100K+ Multi-Tenant Collections
- Support for over 100,000 collections per cluster enables multi-tenant deployments where each tenant maintains isolated vector collections with independent access controls.
- Server-Side Aggregation
- ORDER BY, aggregation functions, and faceted search move result processing from application code into the retrieval engine for reduced network overhead.
- Multi-Language SDK Support
- Official SDKs for Python, Go, Node.js, and Java with Milvus Lite for lightweight local development via pip install and Docker Standalone for single-node deployment.