Milvus

Milvus is a vector database for developers building AI retrieval, image search and other applications that find related content. It stores embeddings alongside metadata for searching text, images and multimodal information.

• Hybrid search: Combine semantic matching with native BM25 full-text search or sparse embeddings such as SPLADE and BGE-M3.
• Filtered retrieval: Restrict similarity searches using metadata stored alongside vectors, including strings, numbers and JSON objects.
• Deployment choices: Use Milvus Lite as a Python library with local file storage, run Standalone on one machine, or use a distributed Kubernetes deployment.
• Growing workloads: Expand read and write capacity independently in distributed deployments, ingest streaming updates and use replicas for fault tolerance.
• Shared storage: Isolate tenants within a cluster and place frequently accessed data in memory or on SSDs, with colder data on slower storage.

Milvus is open source under Apache 2.0. It supports CPU and GPU acceleration, including NVIDIA CAGRA indexing, and index types such as HNSW and DiskANN. Build support includes Linux and both Intel and Apple Silicon Macs.

Self-hosted deployments run on your hardware or servers. Zilliz Cloud is a separate managed offering, with hosted deployments and an option to run in your own cloud account.