SQL Database Vector Search
A Blazor web app and minimal API that performs retrieval-augmented generation and vector search using the native VECTOR type in Azure SQL Database or SQL Server 2025, with Azure OpenAI and Microsoft Agent Framework.
SQL Database Vector Search loads documents in PDF, DOCX, TXT or Markdown format, generates embeddings for them with Azure OpenAI, and stores the vectors using the native VECTOR type in Azure SQL Database or SQL Server 2025, so no separate vector store is required. It performs semantic search and retrieval-augmented generation over the imported documents, accessible either through a Blazor web app or programmatically through a minimal API.
Embeddings and chat completion are orchestrated with Microsoft Agent Framework: an embedding workflow imports documents, a reformulation agent rewrites follow-up questions using conversation context, and a RAG agent is connected to a SQL vector-search context provider.
It targets .NET 10 and requires an Azure OpenAI resource with API keys. It is written in C# and licensed under MIT.