Back to Arena

ParadeDB

by ParadeDB Inc. (YC S23)

System Card

OrganizationParadeDB Inc. (YC S23)
Released2023-12
Architecturevector-rag / Postgres-native full-text and vector search
DetailsParadeDB is a PostgreSQL extension (pg_search + pgvector integration) that brings Elasticsearch-quality BM25 full-text search and hybrid vector search directly inside Postgres. It eliminates the need for a separate search engine by adding inverted index structures natively. Logical replication from existing Postgres instances is supported, requiring no ETL pipeline.
Parameters—
Domainrag-retrieval
Open SourceYes
WebsiteVisit
PostgreSQLBM25hybrid-searchElasticsearch-alternativeextension

Capability Profile

Benchmark Scores

5 of 14 benchmarks
Long-Context Retrieval
2/5
RULER
66.17p
NIAH
no data
LooGLE
no data
LongBench
603p
∞Bench
no data
Multi-Turn Recall
0/2
LoCoMo
no data
MemoryBank
no data
Cross-Session Memory
0/1
LongMemEval
no data
Multi-Hop QA
2/3
BABILong
no data
MultiHop-RAG
65.827p
HotpotQA
60.212p
Agent Task Memory
0/1
AgentBench-Mem
no data
Personalization
0/1
PerLTQA
no data
Factuality / Grounding
1/1
RAGAS
67.842p
Sources:ParadeDB vendor documentation; evaluated on HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering (Stanford / CMU, 1809)ParadeDB vendor documentation; evaluated on LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding (Tsinghua KEG, 2308)ParadeDB vendor documentation; evaluated on MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries (HKUST, 2401)ParadeDB vendor documentation; evaluated on RAGAS: Automated Evaluation of Retrieval-Augmented Generation (Exploding Gradients, 2309)ParadeDB vendor documentation; evaluated on RULER: What's the Real Context Size of Your Long-Context Language Models (NVIDIA, 2404)