Vector

Vector is a list of numbers that represents text or data. In document intelligence, embedding models convert text into vectors of fixed dimensionality. Similar text produces similar vectors; vector search finds content by comparing vectors using similarity score or cosine similarity.

How Vectors Are Used

1. Embed — Text is converted to a vector by an embedding model
2. Store — Vectors are stored in the vector index (Pinecone)
3. Query — Query text is embedded; vector search returns the most similar stored vectors

DocLD uses Pinecone for embedding and vector search. Chunks are embedded and stored; queries are embedded and matched for RAG.

Related Concepts

Vectors are produced by embedding and stored in the vector index. Vector search and similarity score operate on vectors. Dimensionality defines vector size.