A vector database built for billions of embeddings across a cluster of machines.
Real, but not yet worth your team's attention.
Teams genuinely operating at tens of millions of vectors or more, with someone who owns the infrastructure.
A managed vector database bill at serious scale.
Almost every small team that installs this did not need it. Postgres with pgvector handles a few million vectors and you already run Postgres.
Count your embeddings. Under about five million, use pgvector and revisit this in a year.
The numbers
Maintainers describe it as: “Milvus is a high-performance, cloud-native vector database built for scalable vector ANN search”
milvus, in short
- Should a small team use milvus?
- Watch. Real, but not yet worth your team's attention. Teams genuinely operating at tens of millions of vectors or more, with someone who owns the infrastructure.
- What does milvus actually do?
- A vector database built for billions of embeddings across a cluster of machines.
- What does milvus replace?
- A managed vector database bill at serious scale.
- What is the downside of milvus?
- Almost every small team that installs this did not need it. Postgres with pgvector handles a few million vectors and you already run Postgres.
- Can milvus be used in a commercial product?
- Its licence is Apache-2.0, which is permissive and generally fine for commercial use. Confirm against the LICENSE file in the repository.
Weighed against
A memory layer that lets an AI assistant remember facts about a user across separate conversations.
A single application that turns a folder of your documents into a chat assistant, as a desktop app or a hosted install.
Reads text out of scans, photos and PDFs, including tables and forms, and returns structured output.
Self-hosted knowledge graph engine for AI agents