AI

Embeddings

Embeddings are numeric vectors that represent text (or images) so similar meanings sit close together. RAG uses them to find relevant chunks.

In depth

You send text to an embedding API, store the vector, and at query time embed the question and search nearest neighbors (often pgvector).

Cost hits at ingest (every chunk) and sometimes at query. Re-embedding a 10k-doc corpus is a bill, not a footnote.

MVP: one embedding model, one index, metadata filters for tenant_id.

Real example

“refund policy” and “money-back window” retrieve the same chunk even if the words differ.

Tools & calculators

Related terms

Build your product

From definition to shipped MVP in 21 days.

Ready to build your product?

Fixed price. 21-day delivery. Senior team.

Get a free scoping call →

Browse the glossary