Your local RAG isn't slow — it re-reads every document on every question

A 291-second answer over nine files, and what the llama.cpp logs said about where those seconds actually went.