Milvus

Milvus is an open-source vector database for storing and searching embeddings from text, images and other unstructured data.

Milvus is a high-performance vector database written in Go and C++, with CPU and GPU acceleration, under the LF AI & Data Foundation and the Apache 2.0 license. It stores embeddings next to scalars such as integers, strings and JSON, and supports metadata filters and hybrid search. A Kubernetes-native distributed mode handles tens of thousands of queries over billions of vectors with streaming updates; Standalone runs on one machine, and Milvus Lite installs with pip for local work. The Python SDK is pymilvus. Zilliz Cloud is the managed service, with Serverless, Dedicated and bring-your-own-cloud options.

Milvus at a glance

Best for: Software engineer
Runs on: self-hosted, open source, API
Pricing: Free and open source (Apache 2.0); Zilliz Cloud is a paid managed service
Released: September 2019
Category: AI coding tool
Website: github.com/milvus-io/milvus

Key features

Vector search with metadata filters and hybrid search
Distributed Kubernetes-native architecture
Standalone and Milvus Lite for single-machine use
CPU and GPU hardware acceleration
Python SDK (pymilvus) for ingest and query
Managed Zilliz Cloud with Serverless and Dedicated plans
Apache 2.0, hosted by the LF AI & Data Foundation

Use it for

Retrieval for a RAG application over company documents
Image or multimodal similarity search
Local prototyping with Milvus Lite before a cluster

Alternatives to Milvus

zvec — It is a pure local, serverless vector database providing fast hybrid search.
KiviDB — It utilizes every core for maximum performance.
TurboVec — Fits a 10 million document corpus in 4 GB of RAM and searches faster than FAISS.
seekdb — Unifies relational, vector, and text data for production-grade AI agents.
Infinity — Achieves 0.1 milliseconds query latency on million-scale vector datasets.
FastEmbed — It uses ONNX Runtime for high speed and does not require a GPU or PyTorch.
Qdrant — Fast filtered vector search, self-hosted or in a managed cloud
ColdIVF — It reads under 1 percent of the index during cold queries.
Weaviate — Stores data objects and vectors for semantic and hybrid search capabilities.
Singlebase — It combines a document database, vector database, and AI tools in one platform.
HelixDB — Combines graph, vector, and full-text search in one engine.
LlamaIndex — It provides hosted APIs for parsing documents and extracting typed data with one API key.

All Milvus alternatives