Search

Search IconIcon to open search

RAG Indexing Strategies

Last updatedUpdated: by Jakub Žovák · 1 min read

Properties
created 02.08.2026, 10:00
modified 02.08.2026, 10:00
published Empty
sources Empty
topics RAG Indexing, Chunk Preprocessing, Hierarchical Indexing
authors Opus 4.8
ai-assisted Yes
  • RAG methods whose innovation happens at index time — enriching chunks, imposing structure, or precomputing reasoning before any query arrives.

# Topics

  • Contextual Retrieval
    • Prepends LLM-generated, chunk-specific context to each chunk so isolated chunks stay interpretable.
  • RAPTOR
    • Recursively embeds, clusters, and summarizes chunks into a tree, retrieving across abstraction levels.
  • BookRAG
    • Structure-aware index unifying a hierarchical tree with a knowledge graph for complex, hierarchically-organized documents.
  • IndexRAG ( Link)
    • Precomputes multi-hop “bridging facts” at index time so cross-document reasoning needs a single retrieval pass.
  • QB-RAG
    • Pre-generates a database of potential questions from the content and matches incoming queries against it for tighter query-content alignment.