ParadeDB

Postgres-native search engine built on Elasticsearch's BM25 algorithm, implemented as a PostgreSQL extension. ParadeDB adds full-text search (pg_search extension with BM25 ranking), vector search (pg_vectorize), and hybrid search capabilities directly inside PostgreSQL — eliminating the need for a separate Elasticsearch cluster. Search queries run as standard SQL with PostgreSQL's ACID guarantees, joins, and type system. 'Elasticsearch alternative for PostgreSQL users'.

Evaluated Mar 06, 2026 (0d ago) v0.11+
Homepage ↗ Repo ↗ Developer Tools postgresql full-text-search vector-search bm25 hybrid-search open-source postgres-extension
⚙ Agent Friendliness
63
/ 100
Can an agent use this?
🔒 Security
85
/ 100
Is it safe for agents?
⚡ Reliability
69
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
--
Documentation
80
Error Messages
78
Auth Simplicity
90
Rate Limits
90

🔒 Security

TLS Enforcement
95
Auth Strength
82
Scope Granularity
82
Dep. Hygiene
80
Secret Handling
85

AGPL-3.0 — review license for compliance. Standard PostgreSQL security model (TLS, RBAC). Search indexes stored in PostgreSQL — same security boundary as your data. No external data exposure.

⚡ Reliability

Uptime/SLA
72
Version Stability
65
Breaking Changes
60
Error Recovery
78
AF Security Reliability

Best When

You use PostgreSQL and want to eliminate a separate Elasticsearch cluster for search, getting BM25 full-text search and vector search inside your existing database.

Avoid When

You need distributed search at scale, complex search analytics, or are not using PostgreSQL as your primary database.

Use Cases

  • Replace Elasticsearch for full-text search by running BM25 search directly in PostgreSQL without a separate search cluster
  • Implement hybrid search (BM25 + vector similarity) in a single SQL query against your existing PostgreSQL database
  • Build search-enabled applications that need ACID transactions alongside search — Elasticsearch can't JOIN with your Postgres data
  • Add semantic search to existing PostgreSQL applications using pg_vectorize for embedding-based similarity search
  • Implement agent retrieval systems using PostgreSQL's familiar SQL interface for hybrid search without additional infrastructure

Not For

  • Massive-scale search (billions of documents) — Elasticsearch's distributed architecture handles larger workloads than single-node Postgres
  • Complex search features (faceting, geospatial, analytics aggregations at scale) — Elasticsearch has more mature search feature set
  • Teams not already using PostgreSQL — the value is eliminating a separate search cluster; if not using Postgres, use Elasticsearch directly

Interface

REST API
No
GraphQL
No
gRPC
No
MCP Server
No
SDK
No
Webhooks
No

Authentication

Methods: username_password
OAuth: No Scopes: No

PostgreSQL wire protocol authentication — username/password. Standard PostgreSQL RBAC (GRANT/REVOKE). No separate API auth — access via PostgreSQL client connection.

Pricing

Model: open_source
Free tier: Yes
Requires CC: No

AGPL-3.0 — note the copyleft license implications for SaaS deployment (modifications must be open sourced). ParadeDB Cloud is the managed alternative. Self-hosted is free for internal use.

Agent Metadata

Pagination
cursor
Idempotent
Full
Retry Guidance
Not documented

Known Gotchas

  • AGPL-3.0 license — SaaS deployment may require open sourcing modifications; verify license compliance
  • pg_search BM25 index must be created separately from data insertion — agents must run CREATE INDEX USING bm25 before BM25 queries work
  • ParadeDB's BM25 syntax uses @@@ operator — different from PostgreSQL full-text search operators (@@); agents need to know which syntax to use
  • Vector search uses pg_vectorize extension with different API than pgvector — check which extension is installed
  • BM25 search over large tables may be slower than Elasticsearch for high-concurrency scenarios — test performance before production
  • ParadeDB is in active development (v0.x) — API may change between versions
  • Index rebuild required after major data changes — partial index updates are available but full rebuild may be needed for optimal search quality

Alternatives

Full Evaluation Report

Detailed scoring breakdown, competitive positioning, security analysis, and improvement recommendations for ParadeDB.

$99

Scores are editorial opinions as of 2026-03-06.

5215
Packages Evaluated
26151
Need Evaluation
173
Need Re-evaluation
Community Powered