Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
This article lists the various features and improvements that are available in Edge RAG.
Important
Edge RAG Preview, enabled by Azure Arc is currently in PREVIEW. See the Supplemental Terms of Use for Microsoft Azure Previews for legal terms that apply to Azure features that are in beta, preview, or otherwise not yet released into general availability.
July 2025
Release of extension version 0.1.5
The newly released Edge RAG extension version 0.1.5
includes the following changes:
- Resolved an authentication issue for endpoints created by Azure AI Foundry for bring your own model (BYOM) deployments.
- Disabled chat history temporarily while performance improvements are being made for a future release. Each question is answered based on retrieved content only and doesn't include the context of the chat history. Treat each question as a new chat.
- Improved security.
June 2025
New article: Create an endpoint to use for Edge RAG deployment
If you plan to use your own language model instead of one of the models provided by Microsoft, you must set up an OpenAI API compatible endpoint to use with Edge RAG. For more information, see Create an endpoint to use for Edge RAG deployment.
Prerequisites for deployment reorganized into a checklist
To prepare for your deployment of Edge RAG, complete the steps listed in the new checklist: Deployment prerequisites checklist for Edge RAG Preview enabled by Azure Arc. To improve the documentation quality and experience, each deployment prerequisite is now in a separate article.
May 2025
Edge RAG in public preview
Extension version: 0.1.3
Edge RAG is now available as a public preview. To learn more, see the following documentation:
See the related blog posts:
- Transforming On-Premises Data with RAG Capabilities on Azure Local
- Empowering the Physical World with AI - Unlocking AI at the Edge with Azure Arc
- Unlocking AI Apps Across Boundaries with Azure