Embeddings (AI)

What Is Embeddings (AI)?

Embeddings convert human-readable content into vectors (lists of numbers) that capture meaning. Similar concepts have similar embeddings — "dog" and "puppy" are close together, while "dog" and "database" are far apart.

How embeddings work: text goes through an embedding model (like OpenAI text-embedding-3-small or Cohere embed-v3), which outputs a vector of 256-3072 dimensions. These vectors are stored in a vector database for fast similarity search.

Uses: semantic search (find documents by meaning, not keywords), RAG systems, recommendation engines, duplicate detection, clustering, and classification.