Before You Retrieve Anything, Someone Has to Read the PDF
Every retrieval method assumes clean structured text. Getting it out of a PDF is its own problem, and Docling is the 2026 answer worth knowing.
Retrieval and RAG07
Subject · 5 posts
RAG
Several Retrievers, One Small Machine
How to run keyword, dense, fusion and reranking together on modest hardware: where the memory actually goes, and which famous constants are folklore.
Retrieval and RAG06
PageIndex: Retrieval With No Vectors At All
PageIndex indexes a document as its own table of contents and lets a model walk the tree. No embeddings, no chunking, and a route you can show a reader.
Retrieval and RAG04
When Your Knowledge Base Should Be a Graph
GraphRAG turns a corpus into entities, communities and LLM-written reports so it can answer whole-corpus questions that top-k retrieval structurally cannot.
Retrieval and RAG02
Retrieval Metrics Demystified: From BM25 Baselines to EM@5 & Answer F1
What BM25, Recall@k, answer-in-context rate and token F1 actually measure, which number to fix first, and how to compute them without fooling yourself.
Retrieval and RAG01