Fundamentals

Embeddings

An embedding converts text (words, sentences, or documents) into a vector of numbers that captures its semantic meaning. Similar texts produce similar embeddings. Embeddings are fundamental to semantic search, clustering, recommendation systems, and as input to machine learning models. They enable you to work with text numerically. Modern embedding models are trained to capture meaningful relationships—synonyms cluster together, and distance in vector space correlates with semantic distance. Embeddings are typically lower-dimensional (384-1536 dimensions) than raw text.