Vectorize
Vectorize is Cloudflare’s vector database: it stores embeddings and answers similarity queries over them. It is the retrieval half of a retrieval-augmented generation flow — the index a question is matched against before a model is asked to answer it.
What it means
An embedding turns a piece of text into a list of numbers whose distance from other lists approximates meaning. A vector index stores those lists with the metadata needed to find the source again, and a query returns the nearest neighbours of the embedded question along with their scores.
How SwiftTune uses it
The Vectorize block does two things: upsert, which writes vectors and their metadata into an index, and query, which returns the closest matches to an input vector. Indexes are created in the EU jurisdiction where residency is required, and the retrieval monitor watches which documents queries actually return over time.