ParadeDB
activeby ParadeDB, Inc. · since 2023 · also known as pg_search
Elasticsearch-quality BM25 search and fast aggregations inside Postgres via the pg_search extension.
Overview
ParadeDB brings real search-engine indexing into PostgreSQL. Its pg_search extension embeds Tantivy to build BM25 inverted indexes with tokenizers, fuzzy and phrase matching, faceting and columnar 'fast fields' for aggregations, all queried with plain SQL and kept transactionally in sync with your tables. Combined with pgvector it covers hybrid retrieval without running a separate Elasticsearch cluster and an ETL pipeline. The community build is AGPL-3.0; ParadeDB Enterprise adds physical replication of search indexes for HA, and ParadeDB Cloud/BYOC offer managed options.
Strengths & trade-offs
Strengths
- +BM25 full-text search that stays transactionally consistent with your Postgres rows
- +Fast faceting and aggregations via Tantivy columnar fast fields
- +Hybrid keyword plus pgvector retrieval in one SQL query
- +Removes the Elasticsearch sync pipeline for many apps
- +Runs as an extension or as a logical replica beside managed Postgres
Trade-offs
- −AGPL licence and Enterprise-only index replication
- −Not installable on most hyperscaler managed Postgres (RDS, Cloud SQL)
- −Single-node write scaling bounded by Postgres
- −Younger than Elasticsearch for exotic relevance features
Great for
- +Search inside a Postgres-backed SaaS app
- +Hybrid search for RAG next to relational data
- +Faceted catalog and marketplace search
- +Replacing Elasticsearch + ETL with one database
Think twice for
- −Petabyte log analytics
- −Teams locked to RDS/Cloud SQL without logical replicas
- −Massive multi-tenant search clusters
Capabilities
Native or official first-party support · 16 of 25 tracked
Scale & consistency
- Comfortable data size
- Medium · ≤ 10 TB
- Horizontal scaling
- None
- Multi-region
- Read replicas
- Transactions
- Full ACID
- Consistency
- Strong
- Schema
- Strict
- Architecture
- Single node, Primary replica, Client server
- Query languages
- SQL
Scales like Postgres: vertically and with read replicas. Replicating BM25 indexes to physical replicas requires Enterprise; logical replication from an existing primary is supported.
Deployment
available on · Railway · Render · Fly.io · DigitalOcean
Ecosystem
Alternatives to ParadeDB
Lucene-based distributed search and analytics engine for text, vectors, logs and metrics.
paradedb vs elasticsearch →The standard Postgres extension for vector similarity search with HNSW and IVFFlat indexes.
paradedb vs pgvector →The extensible, standards-first open-source relational database.
paradedb vs postgresql →Apache-licensed, foundation-governed fork of Elasticsearch for search, vectors and observability.
paradedb vs opensearch →In-memory, typo-tolerant search engine in C++ with vector, hybrid and conversational search.
paradedb vs typesense →Open-source Postgres backend platform: database, auth, storage, realtime and edge functions in one project.
paradedb vs supabase →ParadeDB FAQ
What is ParadeDB?
ParadeDB is a search engine database by ParadeDB, Inc., first released in 2023. ParadeDB brings real search-engine indexing into PostgreSQL. Its pg_search extension embeds Tantivy to build BM25 inverted indexes with tokenizers, fuzzy and phrase matching, faceting and columnar 'fast fields' for aggregations, all queried with plain SQL and kept transactionally in sync with your tables. Combined with pgvector it covers hybrid retrieval without running a separate Elasticsearch cluster and an ETL pipeline. The community build is AGPL-3.0; ParadeDB Enterprise adds physical replication of search indexes for HA, and ParadeDB Cloud/BYOC offer managed options.
Is ParadeDB open source?
Yes. ParadeDB is open source. License: GNU AGPL v3 (Community); commercial license for Enterprise.
Does ParadeDB support ACID transactions?
Yes. ParadeDB supports full ACID transactions, with strong consistency.
Does ParadeDB scale horizontally?
ParadeDB does not scale horizontally; it runs on a single node, so you scale it up rather than out. It is comfortable at medium scale (≤ 10 TB). It supports cross-region read replicas.
How can I run ParadeDB?
You can self-host it and use the vendor's managed cloud. Managed offerings include ParadeDB Cloud. It is available on Railway, Render, Fly.io and DigitalOcean.
What is ParadeDB best for?
ParadeDB is a strong fit for search inside a Postgres-backed SaaS app; hybrid search for RAG next to relational data; and faceted catalog and marketplace search. Think twice for petabyte log analytics.
What are the best alternatives to ParadeDB?
Common alternatives to ParadeDB include Elasticsearch, pgvector, PostgreSQL, OpenSearch and Typesense. The right choice depends on your workload, scale and hosting; WhichDB's smart match ranks them against your requirements.
How popular is ParadeDB?
ParadeDB ranks #122 of 844 databases on the WhichDB Score (25.1/100), and #10 among search engine databases. The score combines public adoption signals such as package downloads, GitHub activity, job posts and developer discussion.