Milvus is an open-source vector database built to store and search very large numbers of embeddings, for apps such as semantic search, RAG and recommendations.
In plain words
A vector database keeps embeddings, lists of numbers that capture meaning, and quickly finds the ones closest to a query. Milvus stores vectors alongside ordinary fields such as text, numbers and JSON, so searches can also filter on those fields or mix in keyword search. It is designed to scale: storage and computing are separate, and extra machines can be added for heavy reading or writing. It supports many index types, such as HNSW and DiskANN.
A quick example
Imagine a photo site with millions of images. Each image is turned into an embedding and stored in Milvus with its caption and upload date. A user searches for a beach at sunset; the app embeds that text, and Milvus returns the closest images, filtered to those uploaded this year.
Where you'll meet it
- Text and image search.
- Retrieval for RAG applications.
- Recommendation systems.
Related
Sources
- Docsmilvus-io/milvus, Milvus (GitHub) · read 29 Sept 2026
- Docsmilvus-io/milvus-lite, Milvus (GitHub) · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 29 Sept 2026. Spotted an error? Report it.