Definition
Embeddings are numerical vector representations of text (or images, audio) in a high-dimensional space where semantically similar content is placed closer together. Embeddings are the foundation of RAG systems: documents are converted to embeddings and stored in vector databases, enabling semantic search that finds relevant content based on meaning rather than exact keyword matches.
Example
When you ask "How do I deploy to production?", the embedding model finds documents about "deployment guide," "production setup," and "release process" even without matching exact words.
Why it matters
Embeddings are what make semantic search possible. Without them, search would be limited to exact keyword matching, which misses the vast majority of relevant results.