<aside> 🔔 This page is public.

</aside>

Hybrid Search

What is Hybrid Search?

Hybrid search combines traditional keyword-based and semantic search to provide more accurate and context-aware search results.

Users can choose the degree of the hybrid search mix between keyword and semantic search, between 0 and 1.

A Requirement for Semantic Search - Embeddings

To use semantic search and, thus, hybrid search, Meilisearch needs to know the vector representation (embeddings) of the documents and the query.

To manage the embeddings, users now have two options at their disposal:

• Configure Meilisearch to generate said embeddings at indexing time and during the search (for the given query) via proposed sources (openAI or HuggingFace)
• Provide their own embeddings at indexing time and search time for the query.

<aside> 💡 Refer to the Embedders section to learn more about how embedders can be configured.

</aside>

How Does the Hybrid Search Algorithm Work?

Usage Examples

Hybrid Search API Specification

Embedders

Why?