⬢github Python · 36.8K ★ +79 since we first saw it · pushed 22 h ago · MIT
VectifyAI/PageIndex
📑 PageIndex: Document Index for Vectorless, Reasoning-based RAG
PageIndex is a Python library that does retrieval-augmented generation without a vector database. Instead of chunking documents and searching embeddings by similarity, it builds a hierarchical tree index of each document and lets an LLM reason through the tree to find relevant sections — producing traceable, explainable retrieval. A new SDK ships local mode, plus Flash fast indexing and a filesystem layer for million-document corpora.
Why now: It recently released an SDK with local mode and PageIndex Flash (Aug '26), a file-system layer for scaling to millions of documents, and a document-analysis app — driving fresh discussion around the 'vectorless RAG' idea.
Who it is for: Developers building RAG pipelines over long, complex professional documents like financial reports, legal filings, and technical manuals.
agentic-aiagentsaiai-agentscontext-engineeringinformation-retrievalllmragreasoningretrieval
Stars over our 8 snapshots: 36.7K to 36.8K, since 1 h ago.
Where people talked about it
- ⬢github VectifyAI/PageIndex 3 min ago
API: https://socialmediatrends-api.osmike.com/v1/repos/VectifyAI/PageIndex