Glossary

Hybrid retrieval

Hybrid retrieval is a search method that combines keyword matching with meaning-based (vector) search and merges the two rankings into one result list. It finds text that matches exact words and text that matches the idea behind the query.

What hybrid retrieval is

Keyword search finds exact terms, names and codes. Vector search finds passages with similar meaning, even when the wording differs. Hybrid retrieval runs both and merges the results.

The merge step decides which results rise to the top, so each method covers the other's blind spots.

Why it matters

  • Exact names and ids are found reliably by keyword search.
  • Paraphrased questions are found by meaning-based search.
  • Neither signal alone covers every query well.

How it works in Harny

  • You call POST /v1/memory/recall with a query and an identity.
  • Each result carries an id, a rank, occurredAt, content and sources.
  • You set sort to relevance or chronological, and limit from 1 to 200.

Do I need to set up keyword and vector search separately?

No. You call POST /v1/memory/recall with a query and an identity. You do not set up a search index yourself.

Go on. Get Harny.

The first 500 monthly-active end users are free for good. Create an app key and make your first call in minutes.