Vector embeddings

An embedding model maps a piece of text to a vector of numbers — usually several hundred dimensions long. Texts with similar meaning land close together in that space, so distance between vectors approximates similarity of meaning.

AACSearch can generate embeddings server-side (OpenAI, Google, or a built-in model) when you index a document, so your application never handles the model directly, and image embeddings from CLIP enable image search alongside text.