BigHugger
GH Repository · VectifyAI

PageIndex

📑 PageIndex: Document Index for Vectorless, Reasoning-based RAG

stars
35,653
30-day movement
+4415/day
Related entries
60
Connections
2
agent-appagentic-aicontext-engineeringagentsaiinformation-retrievalretrievalragpythonllmreasoningPythonretrieval-augmented-generationai-agentsvector-database

PageIndex is a Python tool from VectifyAI that builds document indexes for vectorless, reasoning-based RAG. Instead of embedding-based retrieval, it structures documents so an LLM can reason over an index to locate relevant content, and it ships a Claude directory (.claude) for agent use.

Reach for it when you want retrieval-augmented generation without maintaining a vector database, letting reasoning replace embeddings for document lookup.

Use it to

  • Index documents for reasoning-based RAG pipelines
  • Replace vector database retrieval with LLM reasoning
  • Integrate document indexing into Claude-based agents
  • Experiment with context engineering for retrieval

For Developers building RAG pipelines and agentic AI applications

Role
agent-app
Language
Python
Licence
MIT
Forks
3,144
Open issues
34
Last push
2026-09-15
Latest release
v0.3.0.dev2 · 2026-07-08
topicsragretrievalreasoningllmagentic-aicontext-engineering