Riven Search

RAG pipelines that actually work in production.

The Riven Knowledge Engine handles the full retrieval-augmented generation pipeline — from document ingestion and chunking to hybrid BM25 + vector search and re-ranking. Connect web crawlers, S3 buckets, databases, and APIs as live knowledge sources.

BM25+
Hybrid Search
< 30ms
Retrieval p99
10M+
Chunks / index
Loading demo…

What's included

Everything you need, nothing you don't.

Hybrid Search

BM25 sparse retrieval combined with dense vector search. Re-ranking with cross-encoders gives you precision without sacrificing recall.

Multi-Source Ingestion

Crawl websites, ingest PDFs, connect S3 buckets, Notion workspaces, GitHub repos, and any REST API with built-in connectors.

Live Knowledge Sync

Set up sync schedules or webhook triggers. Your knowledge graph stays fresh without manual re-indexing.

Retrieval Analytics

Track retrieval quality with MRR, NDCG, and hit-rate metrics. Debug poor retrievals with query-level traces.

Full capability list

BM25 + dense vector hybrid search
Cross-encoder re-ranking
Web crawler with JS rendering
PDF, DOCX, HTML, Markdown ingestion
S3, GCS, Azure Blob connectors
Notion, Confluence, GitHub connectors
Incremental and full re-indexing
Chunk overlap & semantic splitting
Multi-tenant namespace isolation
OpenAI embedding compatibility

Early access

Riven is in beta — pricing opens as we leave beta. Request access and we'll reach out within a few days.

Cookie Preferences

We use essential cookies to operate the site. Optional cookies help us improve your experience. Cookie Policy