Overview

Elasticsearch Search provides advanced hybrid search capabilities that combine vector similarity search with traditional text-based search. This service has proven to be the best-performing overall solution for RAG applications when configured in hybrid mode, creating both text and vector indexes during database creation and leveraging hybrid search for optimal retrieval results.

The hybrid approach excels at handling cases that pure vector search struggles with, such as queries involving specific codes, product IDs, error codes, technical identifiers, and exact terminology matches. By combining semantic understanding through vectors with precise text matching, it delivers superior search quality for real-world knowledge bases.

Service Information

When to Use This Service

Elasticsearch Search is ideal for:

• Production RAG applications requiring the highest search quality
• Hybrid search requirements combining semantic and keyword search
• Technical documentation with codes, IDs, and specific terminology
• Complex knowledge bases with diverse content types and search patterns
• Enterprise search requiring advanced filtering and analytics capabilities
• Best overall performance when search quality is the primary concern

API Requirements

• Elasticsearch Cluster: Access to Elasticsearch cluster with vector search capabilities
• Authentication: Valid credentials for Elasticsearch cluster access
• Hybrid Indexing: Database must be created with both text and vector indexes

Parameters

Search Mode Configuration