Showing archive results for June 2026

AI applications built on Azure Cosmos DB depend on embeddings for grounded results. Keeping them in sync with your data is the hard part: it means building and operating a separate data pipeline to track changes, call an embedding model, and write the results back to Azure Cosmos DB. In practice, that pipeline also has to handle failures and retrie...

Today we are open-sourcing OmniVec, a platform for building and operating the embedding pipelines that keep the vector representation of your operational data in sync as it changes. You register data sources, embedding model(s), vector stores (destination), and OmniVec does the rest: initial backfill, change tracking, model invocation to geenrate, ...

Modern applications don't just write data and move on. They react to it. A new order triggers an inventory update. A profile change syncs to a search index. A deleted record kicks off a compliance workflow. Azure Cosmos DB's change feed makes this possible by giving you a real-time stream of every change happening in your container, and it's one of...

We're excited to announce the general availability of Change Partition Key in Azure Cosmos DB for NoSQL, now with online copy support. You can repartition containers directly in the Azure portal with near-zero downtime and without stopping writes to the source container. Why Partition Key Choice Matters In Azure Cosmos DB, the partition key is on...

Today, we are excited to announce the public preview of the Azure Cosmos DB Migration Assistant for RDBMS to NoSQL, now available in the Azure Cosmos DB extension for Visual Studio Code. 📈 Modernize with confidence Why migrate from RDBMS to Azure Cosmos DB? Modernizing relational workloads has traditionally been complex, time-consuming...

Today, we are excited to announce the General Availability of Per Partition Automatic Failover (PPAF) for Azure Cosmos DB NoSQL API. PPAF is a significant advancement in how Azure Cosmos DB delivers availability and resilience for mission-critical workloads running on single-write-region accounts. If you rely on Azure Cosmos DB to be always on f...

Since we introduced the Azure Cosmos DB MCP Toolkit at Ignite 2025 in preview, the response has been clear: developers want a straightforward way to connect AI agents to their production databases. Customers asked for stability, broader embedding provider support, and a smoother path from experimentation to production. Today, we're announcing th...

The Azure Cosmos DB vNext emulator is generally available today. It ships as a Docker image that runs on Linux, macOS, and Windows, on both x64 and ARM64 architectures, giving you a local Cosmos DB instance you can develop and test against. Use it for inner-loop development on your laptop, in CI integration tests, and anywhere else you’d rather not...

Today we’re thrilled to announce the public preview of Semantic Reranker in Azure Cosmos DB for NoSQL,a new AI-powered capability that improves the relevancy of your search results with just a few lines of code. If you’ve ever run a vector, full-text, or hybrid search and wished the most relevant documents bubbled to the very top, this one’s for yo...

Have you ever needed to evolve your read patterns without losing efficiency? Maybe your app started with one access pattern, but now you need three, and the queries that don't hit your partition key are getting slower and more expensive every month. We built global secondary indexes (GSIs) for exactly this, and today we're thrilled to announce they...