Hybrid Search (Dense + Sparse Retrieval)
Hybrid Search merges dense vector search with sparse keyword matching to deliver superior accuracy across both conceptual and exact-term queries.
Hybrid search is a multi-modal retrieval architecture that combines dense vector embeddings with sparse lexical search (e.g. BM25/tsvector) using rank fusion algorithms.
How Hybrid Search works in practice
Dense semantic search excels at understanding conceptual questions but can fail on exact model numbers, error codes, and alphanumeric SKUs. Lexical search excels at exact keywords but fails on synonyms.
Hybrid search runs both retrieval engines in parallel and merges their candidate lists, ensuring you get both broad semantic comprehension and pin-point keyword accuracy.
How SiteMind implements Hybrid Search
SiteMind runs hybrid search directly in PostgreSQL, combining 768-dim `pgvector` queries with full-text `tsvector` and trigram similarity via Reciprocal Rank Fusion (RRF).
Test our AI tools in your browser (100% Free)
Estimate support savings, token counts, or test prompt injection security guardrails with our zero-cost sandboxes.
Related Technical Concepts
Reciprocal Rank Fusion (RRF)
Reciprocal Rank Fusion (RRF) is an algorithmic scoring method that merges and ranks search results from multiple independent retrieval pipelines.
Full-Text tsvector Indexing
tsvector is PostgreSQL’s native text search type that parses, stems, and indexes words for lightning-fast lexical keyword searching.
Trigram Fuzzy Matching (pg_trgm)
Trigram matching breaks text into 3-character slices, allowing search engines to match queries even when visitors make typos or misspellings.
Vector Database & pgvector
A vector database is specialized storage optimized to index high-dimensional embeddings and execute sub-second approximate nearest neighbor (ANN) searches.
Test it on your own website in under 2 minutes.
Enter your domain to index your pages and preview live answers.