RAG & knowledge

Embedding

A vector of numbers that represents the semantic meaning of a piece of text.

An embedding is a list of floating-point numbers (e.g., 1,024 dimensions) that captures the semantic meaning of a sentence or paragraph. Texts with similar meaning produce similar vectors — even if the words are different.

Embeddings are the backbone of semantic search and RAG: you embed every chunk of your knowledge base ahead of time, then at query time you embed the user's question and find the closest chunks by cosine similarity.

Popular embedding models: OpenAI text-embedding-3-large (3,072-d), Cohere multilingual-v3, Voyage AI, open-source sentence-transformers.

Esempio in GlobalChatbot

GlobalChatbot embeds every paragraph of your uploaded PDFs and crawled site pages. At query time, the bot finds the right answer even if the user phrases it differently than your documents.

Vedi in azione.

GlobalChatbot — agente AI per aziende serie. Configurazione in 5 minuti, 45 lingue, senza carta richiesta.

14 days · no card · cancel anytime