ROCm-RAG documentation

2026-04-28

1 min read time

Applies to Linux

Build and deploy end-to-end AI pipelines with ROCm Retrieval-Augmented Generation (RAG) on AMD GPUs. RAG is a machine learning architecture that enhances Large Language Models by combining generation with information retrieval from external sources.

This documentation demonstrates how you can use RAG for document ingestion, embedding, retrieval, and generation. It outlines the necessary steps and components required to construct a complete RAG pipeline for this workflow.

ROCm-RAG is part of the ROCm-LLMExt toolkit.

The ROCm-RAG public repository is located at ROCm/rocm-rag.

Install

• Install ROCm-RAG

• Configure environment variables

How to

• Run ROCm-RAG in an interactive session

• Run ROCm-RAG with direct execution

• Test the RAG pipelines with Open-WebUI

• Migrate ROCm-RAG Weaviate data

To contribute to the documentation, refer to Contributing to ROCm.

You can find licensing information on the Licensing page.