Skip to content
Glossary

Hybrid Retrieval

Hybrid retrieval combines semantic vector search with lexical keyword search to retrieve relevant passages. Vector search captures meaning and paraphrase, keyword search captures exact terms and identifiers, and a fusion step merges both result sets so neither precise tokens nor conceptual matches are missed.

Synonyms: hybrid search, dense-sparse retrieval, vector plus keyword search, fusion retrieval

Hybrid retrieval acknowledges that no single retrieval signal is enough for production quality. Dense vector embeddings excel at conceptual similarity but blur exact identifiers; sparse lexical methods nail precise tokens but ignore meaning. By running both and fusing their rankings, hybrid retrieval improves recall without sacrificing precision, which matters when an answer must be grounded in exactly the right passage. A reranking pass over the fused candidates then sharpens the final context handed to the model.

Frequently asked questions

Why combine vector and keyword search?
Vector search can miss rare exact terms like SKUs or error codes, while keyword search misses paraphrases. Fusing both recovers the strengths of each and raises recall on real-world queries.
How are the two result sets combined?
A fusion method such as reciprocal rank fusion or a weighted score blend reranks the merged candidates, often followed by a cross-encoder reranker for final precision.