# Hybrid Search

> Hybrid search combines keyword retrieval, such as BM25, with dense embedding retrieval and merges the two result lists, so that a query benefits from both exact-term matching and semantic matching.

- Identifier: PTL-0128
- Category: Retrieval & Tool Use
- Canonical URL: https://protologue.com/t/hybrid-search/
- Also known as: hybrid retrieval, reciprocal rank fusion, RRF

## Description

The lists are usually merged by reciprocal rank fusion, which scores each document by the sum of the reciprocals of its ranks in each list and needs no score calibration between systems. Hybrid retrieval followed by a reranker is a common default for RAG pipelines.

## Related terms

- [BM25](https://protologue.com/t/bm25/)
- [Dense Retrieval](https://protologue.com/t/dense-retrieval/)
- [Reranking](https://protologue.com/t/reranking/)
- [Contextual Retrieval](https://protologue.com/t/contextual-retrieval/)
- [Retrieval-Augmented Generation](https://protologue.com/t/retrieval-augmented-generation/)

## Sources

- Cormack et al. (2009). Reciprocal rank fusion outperforms condorcet and individual rank learning methods. https://doi.org/10.1145/1571941.1572114
- Anthropic (2024). Introducing Contextual Retrieval. https://www.anthropic.com/news/contextual-retrieval

## Cite this entry

Protologue. (2026). Hybrid Search. In Protologue: A Taxonomy of Prompting and LLM Techniques (v1.4.0, PTL-0128). https://protologue.com/t/hybrid-search/

License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)