Skip to content

KET-RAG

A RAG framework with knowledge graph indexing

Visit official site

What KET-RAG is

The framework is described as a retrieval-augmented generation system that joins knowledge graphs with multi-granularity indexing for efficient retrieval and generation.

What you can do with it

  • Select key text with SkeletonRAG
  • Link keywords and text with KeywordRAG
  • Retrieve through entity channels

Getting started

  • Install dependencies with Poetry
  • Run the init command to set up files
  • Build the graph and text indexes

Pros & cons

✓ What we like

  • Indexing cost is stated as reduced
  • Flexible configuration and extension are supported

! What to watch out for

  • Retrieval runs through entity and keyword channels
  • Setup uses Poetry for dependency installation

FAQ

What indexes does it use?

The material names a knowledge graph skeleton and a text-keyword bipartite graph.

How are dependencies installed?

The material states Poetry is used to install project dependencies.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review