ParadeDB
A Postgres search extension that combines full-text and vector retrieval with SQL filters and aggregations, available locally or self-hosted under AGPL-3.0.
ParadeDB adds search to the Postgres database that holds your application data. It's for developers building search or AI retrieval who want to keep queries and data in one database without synchronizing a separate search service.
- Full-text search: Rank matches with tunable BM25 scoring, handle typos and fuzzy matches, highlight results, and search with phrases, proximity or regular expressions.
- Vector and hybrid search: Retrieve results by vector similarity or combine vector retrieval with full-text search.
- Filters and joins: Use standard SQL to combine search results with structured filters, related tables and existing Postgres query logic.
- Facets and aggregations: Group matching records into facets and calculate bucket counts or metrics alongside search results.
ParadeDB runs locally with Docker or as the pg_search extension in self-managed Postgres. Its search index lives alongside your tables, without a sidecar process or external search cluster. Search works with Postgres ACID transactions and read-after-write guarantees.
The project is open source under AGPL-3.0 and has free and paid self-managed offerings. It integrates with Drizzle, Django, SQLAlchemy, Rails and EF Core. Agent skills, an MCP integration and a Cursor plugin support AI development workflows.
Logotyper
