Integrated Embeddings in Azure Cosmos DB: Vectors That Stay in Sync With Your Data (Public Preview)

Azure Cosmos DB now automatically generates and maintains vector embeddings for your items as they are written and updated, so the vectors stored alongside your data always reflect its current state. You configure the source properties to embed, the Microsoft Foundry embedding model to use, and the path where the embeddings are stored, and Azure Cosmos DB takes care of the rest, no separate embedding pipeline to build or operate. This capability, called Integrated Embeddings, is now in Public Preview. In this video, we walk through what Integrated Embeddings is, how it's configured, and an end-to-end demo that uses it to power vector search and a simple RAG agent over a sample dataset in Azure Cosmos DB. Links: - Blog post: https://devblogs.microsoft.com/cosmosdb/announcing-the-public-preview-of-integrated-embeddings-in-azure-cosmos-db - Documentation: https://aka.ms/integrated-embeddings-doc