skills.rest
timescale
timescaleOFFICIAL
Explore all skills by this author →
Website𝕏@timescaledbGitHub

postgres-hybrid-text-search

Combine BM25 keyword search with semantic vector search in PostgreSQL.

1.8k|104|Updated Jul 23, 2025
Search Tags:#PostgreSQL#pgvector#BM25#semantic-search#hybrid-search#RRF
One-click install
npx skills add https://github.com/timescale/pg-aiguide --skill postgres-hybrid-text-search
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: postgres-hybrid-text-search
Source: https://github.com/timescale/pg-aiguide/tree/main/skills/postgres-hybrid-text-search
Command: npx skills add https://github.com/timescale/pg-aiguide --skill postgres-hybrid-text-search

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill enables developers to implement hybrid search by combining BM25 keyword search with semantic vector search to deliver more relevant results in PostgreSQL-driven applications. It helps AI-assisted tooling retrieve both exact-match results and conceptually related documents in a single query flow.

Core Features & Use Cases

  • Hybrid search: merge BM25 keyword relevance with vector-based semantic similarity using Reciprocal Rank Fusion (RRF).
  • Setup guidance: demonstrates enabling pg_textsearch, pgvector, and optional vector indexing methods, plus client-side fusion logic.
  • Use Cases: document search, product catalogs, knowledge bases, code search, and QA systems that require both precise terms and semantic understanding.

Quick Start

Enable the extensions and create a sample documents table with both text and vector columns, then perform BM25 and semantic searches in parallel and fuse results client-side via RRF. Example steps:

  • Enable extensions: CREATE EXTENSION IF NOT EXISTS vector; CREATE EXTENSION IF NOT EXISTS pg_textsearch;
  • Create sample table with id, content, embedding;
  • Create BM25 and HNSW indexes;
  • Run parallel queries and fuse results on the client.

Frequently Asked Questions about postgres-hybrid-text-search

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I combine BM25 keyword search and semantic vector search in PostgreSQL?▼

Hybrid search in PostgreSQL combines BM25 and vector similarity by running both query types in parallel, then fusing results client-side using Reciprocal Rank Fusion (RRF) to merge exact keyword matches with semantic meaning.

What is Reciprocal Rank Fusion (RRF) and how does it work for hybrid search?▼

Reciprocal Rank Fusion (RRF) is a client-side technique that merges ranked result lists from BM25 and vector searches into a single output, weighting documents by their positions in each set to balance keyword and semantic relevance.

How do I set up pg_textsearch and pgvector for document search in PostgreSQL?▼

Setup requires enabling pgvector and pg_textsearch extensions via CREATE EXTENSION, creating a table with text and embedding columns, and building BM25 and HNSW indexes to support parallel keyword and semantic queries.

When should I use PostgreSQL hybrid search instead of just semantic or keyword search?▼

Use hybrid search for document search, product catalogs, and knowledge bases where both exact terms and meaning matter, ensuring precise keyword matches and conceptually related documents are retrieved in a single query flow.

Does PostgreSQL hybrid search require client-side logic to merge BM25 and vector results?▼

Yes, this approach requires client-side logic to execute parallel BM25 and vector queries from PostgreSQL, apply Reciprocal Rank Fusion (RRF) to the retrieved sets, and output the final merged document list.

Related Skills

Explore More Software Engineering Skills→
RepairYourTechRepairYourTech
6

pgvector

Integrate vector similarity search into PostgreSQL using the pgvector extension.

Community
Advanced
spideynolovespideynolove
2

hybrid-search-implementation

Fuse vector similarity with keyword search using RRF or linear fusion.

Community
Advanced
wshobsonwshobson
38.6k

vector-index-tuning

Tune HNSW parameters and quantization strategies for vector index performance.

Community
Advanced
CleanExpoCleanExpo
1

Search Indexer

Implement PostgreSQL full-text search indexing with tsvector and GIN indexes.

Community
Advanced