GPT-RAG Solution Accelerator
An enterprise-grade solution accelerator for building agentic retrieval-augmented generation systems on Microsoft Foundry, with Zero-Trust architecture and infrastructure-as-code deployment templates.
GPT-RAG is a solution accelerator that provides architecture templates and deployment assets for building secure, scalable, enterprise-ready agentic RAG solutions on Microsoft Foundry. Orchestration runs on the Microsoft Agent Framework, and Foundry IQ is the default retrieval backend, blending Blob storage, Azure AI Search, Work IQ, Fabric, SharePoint, OneLake, the web, and MCP sources with permission trimming; a direct Azure AI Search backend remains available as a rollback path.
It is built around a Zero-Trust architecture, so components run in a controlled, isolated environment with tightly governed network access and least-privilege communication between services, together with Responsible AI and end-to-end observability features. Deployment options include Basic, Zero-Trust, and network-isolated topologies, with preflight checks, a jumpbox workflow, and container image builds; a hosted topology serves the orchestrator directly from the Foundry Agent Service without provisioning an orchestrator container or a Cosmos DB.
It is released under the MIT license.