Skip to content

KG_RAG

Retrieve prompt relevant context from a knowledge graph

Visit official site

What KG-RAG is

The framework is a task agnostic retrieval augmented generation approach built on a knowledge graph. It uses SPOKE to pull the smallest context that answers a prompt.

What you can do with it

  • Extract prompt relevant knowledge graph context
  • Add domain specific context to general models
  • Work with GPT and Llama models

Example uses

  • Retrieve P53 gene context for a function question
  • Pull drug mechanism context for a summary
  • Generate a condition report from graph entities

Pros & cons

✓ What we like

  • The framework is described as task agnostic
  • Domain specific context is added to general models

! What to watch out for

  • Context comes from the SPOKE graph
  • Prompts drive what context is extracted

FAQ

What does prompt relevant context mean?

The minimal context needed to respond to a user prompt.

Which models are named?

GPT and Llama models.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review