Definition
A specialized database that indexes high-dimensional vectors (mathematical representations of data), enabling semantic search rather than just keyword matching.
Why It Matters
Traditional databases (SQL) can find 'apple' if you search 'apple'. Vector DBs can find 'apple' if you search 'fruit'. They are the memory backbone of modern AI applications.
How It Works
- 1
Text/Images are passed through an Embedding Model (e.g., OpenAI text-embedding-3).
- 2
The model outputs a vector (e.g., [0.1, 0.5, -0.9...]).
- 3
The Vector DB stores this.
- 4
Queries calculate 'cosine similarity' to find the nearest matches.
The NetForce Take
We use Pinecone or pgvector (Postgres) for almost every AI project. Don't build a custom search engine; use vectors.