RAG :: RETRIEVAL

    What Is Hybrid Search?

    Hybrid search combines two retrieval methods, keyword search and vector search, to return results that match both exact terms and conceptual meaning. Keyword search catches precise matches, while vector search catches conceptually related content even when the wording differs. Combining them, often with a fusion method like reciprocal rank fusion, produces more complete and relevant results than either approach alone.

    How hybrid search works

    A hybrid search system runs a keyword search and a vector search in parallel, then merges the two result sets into a single ranking, commonly using reciprocal rank fusion. The keyword side ensures exact matches are not missed; the vector side ensures conceptually relevant results are included.

    When a separate hybrid pipeline becomes optional

    Hybrid search is usually built as two systems running in parallel, which adds infrastructure and maintenance. Architectures that reconcile lexical and semantic signal within a single retrieval pass can deliver hybrid-quality results without a separate keyword pipeline to operate. Green Vectors takes this approach, reconciling both kinds of signal in one pass.

    More questions

    Often, for mixed queries. Hybrid search adds keyword matching to catch exact terms pure vector search can miss, while keeping vector search's conceptual reach.
    Not necessarily. Some architectures reconcile lexical and semantic signal in a single retrieval pass, delivering hybrid-quality results without operating a separate BM25 pipeline.

    Ready to go deeper?

    Request access to Kitana, our Python SDK built on Green Vectors, or get in touch with the team.