Kition
Download
Researchers

Feeding PDFs to the agent: local RAG in practice

Local PDF extraction + full-text doc index = paper Q&A without going to the cloud.

Kition Team·Jun 12, 2026·1 min read

Flow

  • Drop PDFs into Resources/papers/
  • Agent has a pdf_extract tool: extract body to a .md cache
  • Full-text index covers Notes + papers
  • On question: retrieve → match → answer citing exact paragraphs

When not to RAG

A short paper (<30K tokens) often does better fed in full; RAG shines across dozens / hundreds, not for one deep read.

← Back to blogMore in Researchers →

Related reading

Researchers

A researcher’s literature-review workflow

Search → ingest PDFs → agent summarizes "what this cluster says" → you take the stance.

Files, backup & privacy

Where does Kition keep your data? File layout, explained

Under `<Vault>/`: `.md` files, `.kition/`, `Tables/` — one diagram covers it.

Ready when you are.

Kition is a local-first AI workspace. Markdown documents, structured tables, and an AI agent — running on your own machine, against the model provider you choose.

Download KitionPricing