Guides

Vector / RAG search

Retrieve the most relevant chunks for a natural-language query.

Overview

Retrieve the most relevant chunks for a natural-language query. Spark embeds and indexes your objects, so you drop your separate vector database. This guide covers embedding models, top-k, and chunking.

Before you start

You need a Spark workspace and an API key. See Get started for the prerequisites.

Prototype page

This guide is part of the Spark beta docs. The prototype shows its place in the information architecture. The full step-by-step content lands with the beta.

Was this page helpful?