VectifyAI/PageIndex

Index: generate a tree-structure index for each document
Retrieve: agentically search that tree with LLM reasoning

Maintainer? Show your league in your README

Card: your league's emblem, from Copper to Legendary

[![bestofgithub](https://bestofgithub.dev/badge/VectifyAI/PageIndex.svg?style=card)](https://bestofgithub.dev/repo/VectifyAI/PageIndex)

Flat: fits beside your other badges

[![bestofgithub](https://bestofgithub.dev/badge/VectifyAI/PageIndex.svg)](https://bestofgithub.dev/repo/VectifyAI/PageIndex)

Loading README…

	Vector RAG	PageIndex
Index	vector index	tree index
Retrieval	semantic similarity search	LLM reasoning over the tree
Result	opaque, “vibe retrieval”	traceable to explicit references
Context	query embedding only	full context: conversation history, domain knowledge, etc.

	Local	Cloud
Handles	Text-based PDFs	Text-based, scanned, and image-rich documents
Indexing	On your machine	Managed by PageIndex
Storage	Local directory	Cloud storage
Citations	Page-level	Block-level
OCR & image understanding	—	✓
Metadata	—	✓
Folders	—	✓
MCP server	—	✓

PageIndex: Vectorless, Reasoning-based RAG