Overview

The ideas that make Vectorless different from vector RAG.

Updated 2026

Vectorless rests on four ideas. Read them in order — each builds on the last.

How a document becomes a navigable hierarchy.

Why fixed-size chunks and embeddings are skipped entirely.

The agent strategy that walks the tree to the answer.

Why answers are traceable to exact nodes by construction.

Quickstart

Install an SDK, ingest a document, and ask your first question.

Tree retrieval

How Vectorless turns a document into a navigable hierarchy.