Glossary · Data & retrieval

ETL (extract, transform, load)

The pipeline that takes a source, cleans and shapes it, chunks and embeds it, and persists it into a queryable store.

What is etl (extract, transform, load)?

ETL here is one endpoint that covers the whole path: fetch or accept a source (URL, HTML or raw text), clean it, optionally run structured extraction as the transform step, chunk it, embed the chunks, and load them into a content store scoped to a named collection. Only the structured-extraction step costs model tokens.

Why it matters

A retrieval pipeline over web content is normally four services glued together, each with its own failure mode and bill. Collapsing it into one call removes the glue — and makes the cost of keeping a corpus fresh a number you can actually compute.

Related terms

See it in context

Back to the full glossary

Put your agent to work. Keep the decision.

Give your agent a real browser, authorize the accounts it may use, and keep control of every step. Repeated runs compile into skills that replay without a model.