Page Index

Inspired by AlphaGo, we propose  PageIndex — a vectorless, reasoning-based RAG system that builds a hierarchical tree index for long documents and reasons over that index for retrieval. It simulates how human experts navigate and extract knowledge from complex documents through tree search, enabling LLMs to think and reason their way to the most relevant document sections. It performs retrieval in two steps:

1. Generate a “Table-of-Contents” tree structure index of documents
2. Perform reasoning-based retrieval through tree search
   

Compared to traditional vector-based RAG, PageIndex features:

• No Vector DB: Uses document structure and LLM reasoning for retrieval, instead of vector search.
• No Chunking: Documents are organized into natural sections, not artificial chunks.
• Human-like Retrieval: Simulates how human experts navigate and extract knowledge from complex documents.
• Transparent Retrieval Process: Retrieval based on reasoning — traceable and interpretable. Say goodbye to approximate vector search (“vibe retrieval”).

PageIndex powers a reasoning-based RAG system that achieved  98.7% accuracy on FinanceBench, demonstrating state-of-the-art performance in professional document analysis (see our  blog post for details).