RAG DB Documentation
The indexing engine for enterprise RAG on Azure
RAG DB is a production-grade indexing platform for Retrieval-Augmented Generation (RAG) on Azure. It watches Azure Storage for content changes, processes events through Service Bus, and fans out to index processors that chunk, embed, and store content in Azure Cosmos DB with vector and full-text search.
Sections
- Getting Started — Quick overview and prerequisites
- Architecture — System design, end-to-end flow, infrastructure components
- API Reference — REST API control plane and query gateway
- Search & Queries — Query types, scoring formulas, query rewriting, and performance guidance
- Index Processor — Data-plane worker that creates searchable chunks
- markitdown-pro — Open-source document converter (30+ file types)
- Configuration — All environment variables organized by subsystem
- Security — Authentication, authorization, secrets management, and hardening