ParadeDB
Postgres extensions for search and analytics, an Elastic alternative.
About
ParadeDB is an open-source Postgres extension stack that turns PostgreSQL into a real-time search and analytics engine, positioned as an Elasticsearch alternative without leaving Postgres. Its headline extension is pg_search, which brings BM25-ranked full-text search into Postgres via the Tantivy engine, queried with the @@@ operator and paradedb.score() and paradedb.snippet() functions.
An earlier companion extension, pg_analytics, put DuckDB inside Postgres for OLAP over object stores like S3, but it was deprecated in April 2025 as analytics work consolidated into pg_search. ParadeDB ships as a Postgres-compatible Docker image and as standalone extensions, and its hybrid search with vector support makes it a natural fit for RAG workloads that want search, vectors, and aggregations in one database.
What it can do
- BM25 full-text search inside Postgres via pg_search
- Hybrid text, vector, and aggregation queries in one index
- Postgres-compatible wire protocol with SQL queries
- Pure extension: no separate search server or ETL sync
Setup
Run the paradedb/paradedb Docker image or install the pg_search extension on an existing Postgres instance.
Prerequisites
Postgres 14+ (13 deprecated) and Docker for the prebuilt image.
Use cases
Works with
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 95; Trending score 76/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.