Microsoft announces preview of data ingestion building blocks for .NET, providing modular abstractions for building AI data pipelines. The library offers unified document representation, flexible readers for multiple file formats, built-in AI enrichments (summarization, sentiment analysis), customizable chunking strategies, and integration with vector databases. Built on Microsoft.Extensions.AI and Microsoft.Extensions.VectorData, it enables developers to create composable ETL workflows for RAG scenarios with support for various storage backends including Qdrant, SQL Server, and MongoDB.
Table of contents
What is data ingestion? Copy linkIntroducing Data Ingestion Building Blocks in .NET Copy linkBuild your first data ingestion pipeline Copy linkEnd-to-End Scenarios Copy linkReady to get started? Copy linkSort: