PageIndex
Vectorless RAG system that navigates a hierarchical document tree with LLM reasoning.
About
Skipping embeddings entirely, PageIndex retrieves the way a person uses a table of contents: Vectify AI's MIT-licensed system parses a document into a hierarchical tree index, then lets an LLM navigate that tree with multi-step reasoning and tree search to locate relevant sections, with no vector database, chunking strategy, or similarity threshold anywhere in the pipeline. The approach suits long, structured material like financial reports, legal filings, and technical manuals, where the team reports 98.7 percent accuracy on the FinanceBench benchmark. A pip install pageindex gets the open source SDK, which runs in local mode against a caller-supplied LLM API key or in cloud mode where hosted OCR, tree building, and retrieval sit behind a PageIndex key from the developer dashboard. Native integrations cover OpenAI, Anthropic, the Claude Agent SDK, LangChain, and PydanticAI, and an MCP server exposes retrieval to Claude and Cursor. Documentation lives at docs.pageindex.ai. Around 35,000 GitHub stars later, it anchors the vectorless side of the ongoing RAG architecture debate.
Should you use PageIndex?
Pick it when
Choose PageIndex for long, structured documents such as financial filings, contracts, or technical manuals, where answers depend on the section hierarchy and chunk similarity search keeps retrieving the wrong passages.
Look elsewhere when
Skip it for large corpora of many short documents or when per-query LLM cost and latency matter, since retrieval runs multi-step LLM reasoning over the tree. A vector store such as Qdrant plus a reranker is cheaper at scale.
Alternatives to PageIndex
- RAGFlow
Template-based chunking with citations across many documents, scales to large collections more cheaply, but relies on chunks and embeddings.
- LightRAG
Graph plus vector retrieval across a whole corpus, for questions that span many documents rather than drilling into one long file.
- Rerankers
Adds a reranking stage to ordinary vector search, a cheaper way to fix poor chunk ranking without LLM tree navigation on every query.
- Morphik
Searches pages visually, better for charts and tables, while PageIndex reasons over text structure; Morphik is BUSL-1.1 where PageIndex is MIT.
Reviews (0)
Leave a Review
No reviews yet. Be the first to review!
Details
- Category
- RAG & Document Retrieval
- Price
- Freemium
- Platform
- Hybrid
- Difficulty
- Intermediate (3/5)
- License
- MIT
- Added
- Aug 24, 2026