Meaning Vectors by API - Embeddings and Text Enrichment

The vectors behind search by meaning, for your own code

Meaning vectors

A meaning vector is what lets a search for "cheap flights" find a page about "low-cost airfare". Opensolr makes one for every crawled and pushed document on its own; these calls give the same vectors to your own code, made by the same model, so they match the ones in your index.

Every call here goes to https://api.opensolr.com. Each text turned into a vector counts against the monthly AI allowance of the account that owns the index; a vector already in cache costs nothing (API Quota).

The calls

embed

One text, or one search question, turned into a meaning vector.

Reference
batch_embed

Many texts turned into vectors in one call; each text counts as one request.

Reference
embed_opensolr_index

Add meaning vectors to the documents already in an index.

Reference
enrich_text

The language, locale, sentiment and, when you ask, the vector of a list of texts, worked out exactly as for a pushed document.

How vectors turn into better results: Meaning vectors, done for you, Hybrid search and What Opensolr adds to every document.

Site search API pages

Back to Enterprise Site Search