paradedb/paradedb

Health v2 · maintenance only

Adoption not part of the score

9199 stars · 440 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

ParadeDB is a Postgres extension (pg_search) that adds a custom index for fast full-text search, vector search, BM25 scoring, filtering, and aggregations directly inside PostgreSQL. It lets teams run application data and search workloads in one database instead of syncing to an external engine like Elasticsearch.

Use cases

• add full-text search to my postgres database
• replace elasticsearch with postgres
• run vector similarity search in postgres
• hybrid keyword and semantic search in one database
• bm25 relevance scoring with sql
• avoid syncing postgres data to a separate search engine
• real-time analytics and aggregations on postgres
• build rag retrieval over application data

When to choose

• Postgres is already your primary database and you want search without a second system
• You've outgrown tsvector or pgvector performance or features
• You want to eliminate ETL pipelines syncing data to Elasticsearch
• You need BM25, vector, and hybrid search with transactional guarantees in standard SQL

When to avoid

• You don't use PostgreSQL as your database
• You need a standalone search engine decoupled from your OLTP database
• You require managed cloud search today (ParadeDB Cloud is still in waitlist)
• You need Postgres versions older than 15 or unsupported platforms without building from source

Facets

plugin · maturity active

search-engine vector-database database database-driver analytics databases analytics developer-tools windows self-hosted cloud rust cpp postgresql-extension pg_search bm25 full-text-search hybrid-search vector-search columnar-storage htap elasticsearch-alternative sql search retrieval-augmented-generation linux macos docker

Member repositories

For agents

markdown · JSON · MCP: product_card(name="paradedb/paradedb")