Vector database
Level 3
A database built to store embeddings and find the closest matches fast.
Term 4 of 7 in Data and databases
In plain language
A vector database holds those coordinate lists and answers a different question from a normal database: not "which row equals this" but "which items mean something similar to this".
Think of it like this
A librarian who can instantly hand you the ten books nearest to the one you liked.
Why it matters
Vector databases are the standard way to give an AI system access to your own documents.
Next terms in Data and databases
- DatasetA collection of examples gathered for a model to learn from or be tested on.
- Structured and unstructured dataStructured data lives in neat rows and columns; unstructured data does not.
- EmbeddingA list of numbers that captures the meaning of a piece of text or an image.
- Retrieval augmented generationFetching relevant documents first, then asking the model to answer using them.