Intelligent-Internet

Intelligent-Internet / psql_bm25s

Public

PostgreSQL BM25S extension

49
2
100% credibility
Found May 14, 2026 at 80 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
PLpgSQL
AI Summary

psql_bm25s adds high-performance full-text search indexes to PostgreSQL databases using BM25 scoring, optimized for dynamic workloads with automatic maintenance and hybrid fusion support.

How It Works

1
🔍 Need faster text search?

You have a database full of documents and want to find relevant ones quickly using simple keywords.

2
📥 Grab the ready-to-use package

Download the simple zip file or Docker image that matches your database setup from the releases page.

3
⚙️ Add it to your database

Copy the files into place and restart your database, then enable it with one easy command.

4
📝 Tag your text columns

Pick the columns holding your text or lists of words and create a special fast-search index on them.

5
✨ Search and see magic

Type a query like 'apple fruit' and instantly get the best matching documents ranked by relevance.

6
🔄 Keep adding or changing data

Your searches stay fresh automatically as you insert, update, or delete rows—no manual rebuilds needed.

🚀 Lightning-fast searches forever

Enjoy accurate, speedy results even on huge changing collections, blending keywords with other rankings if needed.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 80 to 49 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is psql_bm25s?

psql_bm25s is a PostgreSQL extension delivering BM25-family lexical search indexes directly in your database. Create indexes on text, varchar, or array columns with `CREATE INDEX USING psql_bm25s`, then query via functions like `psql_bm25s_query` or operators such as `@@` and `<=>` for top-k results. Built in PLpgSQL with C backing, it handles mutable workloads—inserts, updates, deletes—with automatic maintenance, plus Docker images via postgresql github docker and prebuilt packages in postgresql releases github.

Why is it gaining traction?

It crushes the Python bm25s reference in BEIR QPS benchmarks (median 4x faster on ids path), while matching semantics and adding Postgres-native durability, replication, and hybrid BM25+vector fusion without pgvector dependency. No more offloading search to external services; queries fuse multi-index results or blend with vectors in pure SQL. Ties seamlessly into postgresql github action CI and postgresql github copilot workflows.

Who should use this?

RAG backend engineers indexing dynamic docs in Postgres, ditching Elasticsearch setup pain. Search-heavy apps on postgresql github repo stacks needing lexical reranking over vectors. Teams evaluating postgresql github mirror or jdbc integrations for production text retrieval.

Verdict

Solid pick for fast, exact BM25 in Postgres—run the postgresql github docker image and benchmarks first. At 49 stars and 1.0% credibility score, it's early but battle-tested with extensive scripts, docs, and replication smoke tests; watch for broader tokenizer support.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.