Definition
A vector database is a storage system purpose-built to hold embedding vectors and answer the question "which stored items are most similar to this one?" Where a traditional relational database matches exact values in rows and columns, a vector database compares the geometric proximity of high-dimensional vectors, returning the items whose meaning is closest to a query.
Similarity Search
The core operation is nearest-neighbor search: given a query vector, find the closest stored vectors using a distance metric such as cosine similarity or Euclidean distance. Because comparing a query against millions of vectors exhaustively is slow, vector databases use Approximate Nearest Neighbor (ANN) indexes, most commonly HNSW graphs, which skip the vast majority of candidates while still returning results nearly identical to an exact search, at a fraction of the cost.
Role in Private AI
Stored vectors typically run from a few hundred to a few thousand dimensions, depending on the embedding model used. For sovereign operators, an open-source vector database (such as those that run entirely on local hardware) lets you build a searchable, private corpus, the memory layer for retrieval-augmented generation, without sending a single document to a cloud provider. Your data, your index, your machine.
A vector database stores the output of an Embeddings model and supplies the retrieved context that a Local LLM reads at query time.
In Simple Terms
A vector database is a storage system purpose-built to hold embedding vectors and answer the question “which stored items are most similar to this one?”…