Glossary

Vector database

A vector database answers one question extremely fast: which of these million vectors are closest to this one.

Start freeNo credit card. 2 minute setup.

Why it matters

Comparing a query against every stored vector works fine for a hundred documents and becomes unusable at a hundred thousand. Vector databases exist to make nearest-neighbour search fast enough to sit in a request.

  • Naive similarity search that scales linearly with content size
  • Retrieval latency that grows as documentation grows
  • No efficient way to filter by metadata and similarity together

What it actually does

It stores vectors with their metadata and indexes them so that approximate nearest-neighbour queries return in milliseconds. 'Approximate' is the key word — it trades a small amount of recall for a very large amount of speed.

Filtering matters as much as similarity

In a multi-tenant product, a vector search that could return another customer's passage would be a serious defect. Filtering by tenant and source alongside the similarity query is what keeps retrieval both relevant and correctly scoped.

Common questions

Do I need to run one myself?
Not to use a product built on this architecture — it is infrastructure, not a feature you configure. It matters to you only in that it is why retrieval stays fast as your content grows.
Is a vector database a replacement for a normal database?
No. It answers similarity queries; it is not where you keep your orders. Most systems run both.

From the blog

All posts

Answer honestly. Capture the rest.

Point Matter Chat at your site and see what it can — and can't — answer. It's honest about both.

Start free — chat in your site

No credit card. 2 minute setup.

Every answer cites the source it came from. When there isn't one, it says so — and hands the visitor to you.

Installs on the tools you already run.