Hybrid RAG Search
Combines full-text and semantic vector search with RRF for unmatched accuracy and relevance.
AI Chat Interface
Conversational AI powered by OpenAI with streaming responses and context-aware retrieval.
Intelligent Search
Natural language search with instant AI summaries of the most relevant knowledge base content.
Knowledge Management
Upload PDFs, ingest URLs, or add manual entries with automatic vectorization.
Lightning Fast
Sub-200ms latency with PostgreSQL + pgvector, deployed on Vercel with Supabase backend.
Multi-Tenant SaaS
Fully isolated tenant data with category-based organization and embeddable widgets.
How Hybrid RAG Works
The power of combining full-text search with semantic understanding
User Query
Full-Text Rank
Vector Rank
RRF Fusion
AI Response
The RRF Formula
Documents ranking highly in both full-text and semantic search receive the highest scores, combining precision (exact matches) with recall (conceptual understanding).
| Method | Precision | Recall | F1 Score | Latency |
|---|---|---|---|---|
| Full-Text Only | 0.72 | 0.58 | 0.64 | 180ms |
| Vector Only | 0.68 | 0.81 | 0.74 | 220ms |
| Hybrid RRF (Cortex) | 0.82 | 0.89 | 0.85 | 195ms |
Built with Modern Tech
Enterprise-grade technologies powering intelligent knowledge management