Microsoft Makes Azure Cosmos DB MCP Toolkit Generally Available

Microsoft announced that the Azure Cosmos DB MCP Toolkit v1.1.2 is now generally available, positioning it as a supported way to connect AI applications to data stored in Azure Cosmos DB for NoSQL through the Model Context Protocol.

What Microsoft released

In a post on the Microsoft Azure Cosmos DB Blog, Microsoft said the Azure Cosmos DB MCP Toolkit is now generally available and described the release as a way to bring Cosmos DB data to AI agents “at scale.” The company’s separate Microsoft Build 2026 Cosmos DB roundup also lists the toolkit as generally available and says it exposes Azure Cosmos DB data to MCP-compatible AI clients and agent frameworks.

Microsoft Learn describes the AzureCosmosDB/MCPToolKit as an open-source solution for secure AI-agent interaction with Azure Cosmos DB through MCP. The documentation focuses on Azure Cosmos DB for NoSQL and includes guidance for Microsoft Foundry integration and production deployment.

The Model Context Protocol, or MCP, is used here as a standardized interface between AI clients and external tools or data systems. In practical terms, Microsoft is offering a Cosmos DB-specific MCP server so compatible AI applications can query or retrieve data from Cosmos DB without each application needing a separate custom connector.

Key capabilities in version 1.1.2

According to the Microsoft Azure Cosmos DB Blog announcement, version 1.1.2 adds multi-provider embedding support. That matters for retrieval-augmented generation and vector search scenarios because organizations may generate embeddings through different providers while storing and querying vectors in Cosmos DB.

Microsoft also says the release adds MCP HTTP transport compatibility. That expands deployment options beyond local development patterns and is relevant for teams that want to run MCP services in more production-oriented environments.

The announcement also highlights Microsoft Foundry integration. Microsoft Learn’s documentation likewise describes integration with Microsoft Foundry, giving developers a path to use the toolkit alongside Microsoft’s AI development environment.

Another named addition is a vector_search tool. Microsoft’s source material describes this as part of the toolkit’s support for connecting AI applications to production vector database data in Cosmos DB. The company’s broader Build 2026 Cosmos DB roundup places the MCP Toolkit alongside other database features such as semantic reranking and global secondary indexes, underscoring Microsoft’s focus on AI retrieval and search workloads in Cosmos DB.

Security and production use

Microsoft’s announcement emphasizes enterprise role-based access control, or RBAC, for the toolkit. Microsoft Learn also frames the MCP toolkit as a secure way for AI agents to interact with Azure Cosmos DB. RBAC support is important because an AI application connected to a database should not automatically receive broad access to every container, item, or operation.

For companies experimenting with AI assistants over internal data, this distinction is significant. A demo can often rely on simplified credentials and local access. A production deployment generally needs narrower permissions, auditable access patterns, and compatibility with existing cloud security practices. Microsoft’s documentation presents the toolkit in that production context rather than only as a developer preview.

Why it matters

The general availability label means Microsoft is moving the Azure Cosmos DB MCP Toolkit beyond an experimental status, according to its Azure Cosmos DB Blog and Build 2026 roundup. For developers already using Cosmos DB for NoSQL, the toolkit offers a Microsoft-provided route to make application data available to MCP-compatible AI clients and frameworks.

The release also reflects a broader shift in AI application architecture. Instead of building one-off integrations for every model, assistant, or workflow, developers are increasingly using protocol-based connectors to let AI systems call tools and retrieve relevant data. Microsoft’s Cosmos DB MCP Toolkit applies that pattern to a managed cloud database that already supports NoSQL data and vector-oriented workloads.

The announcement does not, by itself, prove that MCP will become the dominant integration layer for enterprise AI applications. It does show, however, that Microsoft is treating MCP support for Cosmos DB as a production feature, with attention to embeddings, vector search, HTTP transport, Microsoft Foundry integration, and RBAC.