Search foundation

Embeddings

The first stage of search. An embedding model turns documents and queries into vectors, so nearest-neighbour search finds what a query means — not only the words it happens to use.

In development

What are embeddings?

An embedding is a list of numbers that represents a piece of text. Texts with similar meaning get vectors that sit close together, so a question and the passage that answers it can be matched even when they share few words.

Documents are embedded once and stored in a vector index. At query time only the query is embedded, and nearest-neighbour search returns the closest documents — the candidate set a reranker then refines.

How it works

Embeddings in 3 steps.

  1. 01

    Index

    Embed your documents once and store the vectors in a vector index.

  2. 02

    Query

    Embed each incoming query with the same model.

  3. 03

    Retrieve

    Nearest-neighbour search returns the documents closest in meaning, ready for reranking.

Models

Embeddings models

No Embeddings model is released yet. Models appear here as soon as they are published on Hugging Face.

Compare

Embeddings and BM25

EmbeddingsBM25
Matches onMeaningShared terms
Paraphrases and synonymsFoundMissed unless the query is expanded
Exact names, codes and numbersCan blur themPrecise
IndexVector indexInverted index

FAQ

Embeddings, explained.

Do I still need a reranker?
Embeddings are fast because documents are encoded before any query arrives. A reranker reads the query and the candidates together, which is slower but more precise — so the two are usually combined: embeddings to retrieve, a reranker to order.
Should I combine embeddings with BM25?
Often, yes. Hybrid retrieval merges both candidate lists: embeddings catch paraphrases, BM25 catches exact terms such as names, codes and article numbers.
How do I get access?
Embeddings is in development. To hear when it is released, get in touch.

Start a conversation

Put Embeddings to work
on your search.

Tell us what you retrieve and how you measure quality.

Get in touch →