GitHub - VectifyAI/PageIndex: 📑 PageIndex: Document Index for Vectorless, Reasoning-based RAG
a month ago
- #AI
- #RAG
- #Document Analysis
- PageIndex is a reasoning-based RAG system that eliminates the need for vector databases and chunking by using hierarchical tree indexing.
- It simulates human-like retrieval by organizing documents into natural sections and using LLMs for reasoning-based search.
- PageIndex achieved 98.7% accuracy on FinanceBench, outperforming traditional vector-based RAG systems.
- The system is available as a chat platform, API, or MCP integration, with options for self-hosting, cloud service, or enterprise deployment.
- PageIndex can process lengthy PDFs into semantic tree structures, ideal for financial reports, regulatory filings, and academic textbooks.
- Mafin 2.5, powered by PageIndex, excels in financial document analysis with high precision in navigating complex reports.