{"id":13231,"date":"2026-11-17T10:45:56","date_gmt":"2026-11-17T18:45:56","guid":{"rendered":"https:\/\/devblogs.microsoft.com\/cosmosdb\/?p=13231"},"modified":"2026-11-16T09:07:27","modified_gmt":"2026-11-16T17:07:27","slug":"azure-cosmos-db-ai-agent-developer-tools","status":"publish","type":"post","link":"https:\/\/devblogs.microsoft.com\/cosmosdb\/azure-cosmos-db-ai-agent-developer-tools\/","title":{"rendered":"Accelerate development and AI agent integration with Azure Cosmos DB"},"content":{"rendered":"<p>Building applications with Azure Cosmos DB involves more than writing queries. Developers need to make data-modeling decisions, understand how their workloads behave, and connect application data to emerging AI experiences. They also need the flexibility to work in their preferred environment, whether that is an editor, a terminal, or an AI client.<\/p>\n<p>At Microsoft Ignite, we are announcing updates that support each of these workflows:<\/p>\n<ul>\n<li><strong>AI-assisted data modeler and account dashboard in the Azure Cosmos DB extension for VS Code, in preview<\/strong><\/li>\n<li><strong>Azure Cosmos DB Shell, now generally available<\/strong><\/li>\n<li><strong>Managed MCP Server in Azure Cosmos DB, in preview<\/strong><\/li>\n<\/ul>\n<p>Together, these updates help teams move from designing their data to exploring their accounts and integrating AI agents, with tools that fit into their development workflow.<\/p>\n<h1>Design and explore in VS Code<\/h1>\n<p>The Azure Cosmos DB extension brings database development into the editor where developers already write and debug their applications. With the preview of the AI-assisted data modeler and account dashboard, we are extending that experience to support data-design discussions and account-level investigation.<\/p>\n<h2>Bring AI assistance to data modeling<\/h2>\n<p>Data modeling is an important part of building with Azure Cosmos DB. Decisions about document structure, relationships, and partitioning should reflect how an application reads and writes data, as well as how its workload is expected to grow.<\/p>\n<p>The <strong>AI-assisted data modeler<\/strong> brings AI assistance into this design workflow in VS Code, helping developers explore modeling decisions and consider their trade-offs.<\/p>\n<p>For example, a team building an order-management application needs to think about how customers retrieve orders, how order details are represented, and how traffic is distributed. These are application-design questions, not just schema questions. Bringing AI assistance into the modeling process gives developers another way to work through those decisions while staying close to their application code.<\/p>\n<p>AI assistance complements engineering judgment. Developers should review recommendations and validate their design against representative data and workload requirements before adopting it in production.<\/p>\n<p><a href=\"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-content\/uploads\/sites\/52\/2026\/11\/ai-data-modeler.png\"><img decoding=\"async\" class=\"size-full wp-image-13232 aligncenter\" src=\"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-content\/uploads\/sites\/52\/2026\/11\/ai-data-modeler.png\" alt=\"AI assisted data modeler\" width=\"643\" height=\"322\" srcset=\"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-content\/uploads\/sites\/52\/2026\/11\/ai-data-modeler.png 643w, https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-content\/uploads\/sites\/52\/2026\/11\/ai-data-modeler-300x150.png 300w\" sizes=\"(max-width: 643px) 100vw, 643px\" \/><\/a><\/p>\n<h2>Understand your account with the account dashboard<\/h2>\n<p>Once an application is running, developers need to understand what is happening in their database account.<\/p>\n<p>The <strong>account dashboard<\/strong> brings account information and operational signals into VS Code. Developers can explore databases and containers, review throughput and request unit (RU) consumption, investigate partition health, and examine available alerts, recommendations, and advisories.<\/p>\n<p>This helps connect an observation to an investigation. When RU consumption increases, for example, a developer can look at workload trends and partition behavior before deciding whether to investigate application access patterns or capacity.<\/p>\n<p>Dashboard information depends on the account&#8217;s configuration, available monitoring data, and the user&#8217;s permissions. Recommendations and advisories provide starting points for investigation rather than automatically changing resources.<\/p>\n<p><a href=\"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-content\/uploads\/sites\/52\/2026\/11\/account-dashboard.png\"><img decoding=\"async\" class=\"size-full wp-image-13233 aligncenter\" src=\"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-content\/uploads\/sites\/52\/2026\/11\/account-dashboard.png\" alt=\"Account dashboard in vs code\" width=\"643\" height=\"342\" srcset=\"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-content\/uploads\/sites\/52\/2026\/11\/account-dashboard.png 643w, https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-content\/uploads\/sites\/52\/2026\/11\/account-dashboard-300x160.png 300w\" sizes=\"(max-width: 643px) 100vw, 643px\" \/><\/a><\/p>\n<h1>Explore and automate with Azure Cosmos DB Shell<\/h1>\n<p>For developers who prefer working in a terminal, <strong>Azure Cosmos DB Shell is now generally available<\/strong>.<\/p>\n<p>The shell provides a command-line experience for navigating and working with Azure Cosmos DB resources. Familiar commands such as ls, cd, and pwd help developers move through databases and containers, while database and document operations support everyday development tasks.<\/p>\n<p>Developers can run queries, inspect JSON results, and use pipes and scripts to build repeatable workflows. A sequence that begins as an interactive exploration can become a reusable script for a recurring task.<\/p>\n<p>The shell supports authentication options including Microsoft Entra ID, managed identities, and account keys. Teams can choose an appropriate authentication approach for their environment while keeping credentials out of scripts and applying the permissions required for the task.<\/p>\n<h1>Connect AI agents with Managed MCP Server<\/h1>\n<p>AI applications increasingly need to interact with external systems, not just generate responses from a model&#8217;s existing knowledge.<\/p>\n<p>The <strong>Managed MCP Server in Azure Cosmos DB<\/strong>, now in preview, provides a managed integration point for connecting compatible AI clients and agents to Cosmos DB through the <strong>Model Context Protocol (MCP)<\/strong>.<\/p>\n<p>MCP provides a standard way for AI clients to discover and invoke tools. A managed server experience reduces the need for teams to operate their own MCP server infrastructure as they explore agent integration.<\/p>\n<p>The distinction between an AI-generated answer and a tool-assisted interaction matters: a connected client can use the tools exposed by the server and incorporate their results into its response.<\/p>\n<p>As with any integration that accesses application data, teams should evaluate supported operations, authentication requirements, permissions, and data-handling expectations. An agent&#8217;s natural-language request is not a substitute for authorization.<\/p>\n<p><a href=\"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-content\/uploads\/sites\/52\/2026\/11\/hosted-mcp.png\"><img decoding=\"async\" class=\"size-full wp-image-13235 aligncenter\" src=\"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-content\/uploads\/sites\/52\/2026\/11\/hosted-mcp.png\" alt=\"Azure Cosmos DB MCP server\" width=\"643\" height=\"322\" srcset=\"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-content\/uploads\/sites\/52\/2026\/11\/hosted-mcp.png 643w, https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-content\/uploads\/sites\/52\/2026\/11\/hosted-mcp-300x150.png 300w\" sizes=\"(max-width: 643px) 100vw, 643px\" \/><\/a><\/p>\n<h1>Get started<\/h1>\n<p>Choose the experience that fits your next task:<\/p>\n<ul>\n<li><strong>Design your data:<\/strong> Explore the <a href=\"https:\/\/learn.microsoft.com\/azure\/cosmos-db\/vscode-extension\/data-modeler\">AI-assisted data modeler<\/a> in the Azure Cosmos DB extension for VS Code.<\/li>\n<li><strong>Investigate your account:<\/strong> Learn about the <a href=\"https:\/\/learn.microsoft.com\/azure\/cosmos-db\/vscode-extension\/account-dashboard\">account dashboard<\/a>.<\/li>\n<li><strong>Work from the terminal:<\/strong> Follow the <a href=\"https:\/\/learn.microsoft.com\/azure\/cosmos-db\/shell\/get-started\">Azure Cosmos DB Shell quick start<\/a>.<\/li>\n<li><strong>Connect an AI client or agent:<\/strong> Get started with <a href=\"https:\/\/learn.microsoft.com\/azure\/cosmos-db\/gen-ai\/hosted-mcp-server\">Managed MCP Server in Azure Cosmos DB<\/a>.<\/li>\n<\/ul>\n<p>From AI-assisted design to terminal workflows and agent integration, these announcements give developers more ways to work with Azure Cosmos DB. Start with one task, validate the experience against your needs, and share feedback as you explore the previews.<\/p>\n<div class=\"markdown-heading\" dir=\"auto\">\n<h2 class=\"heading-element\" dir=\"auto\" tabindex=\"-1\"><span style=\"font-size: 16px;\">About Azure Cosmos DB<\/span><\/h2>\n<\/div>\n<p>Azure Cosmos DB is a fully managed NoSQL and vector database for modern applications, including AI apps. With SLA-backed performance and availability, automatic scaling, and global distribution, it helps developers build fast, reliable applications.<\/p>\n<p>Find more developer resources at the <a href=\"https:\/\/cosmosdb.microsoft.com\">Azure Cosmos DB developer hub<\/a>.<\/p>\n<p>Stay connected with Azure Cosmos DB on <a href=\"https:\/\/twitter.com\/AzureCosmosDB\">X<\/a>, <a href=\"https:\/\/aka.ms\/AzureCosmosDBYouTube\">YouTube<\/a>, and <a href=\"https:\/\/www.linkedin.com\/company\/azure-cosmos-db\/\">LinkedIn<\/a>.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Building applications with Azure Cosmos DB involves more than writing queries. Developers need to make data-modeling decisions, understand how their workloads behave, and connect application data to emerging AI experiences. They also need the flexibility to work in their preferred environment, whether that is an editor, a terminal, or an AI client. At Microsoft Ignite, [&hellip;]<\/p>\n","protected":false},"author":80443,"featured_media":13236,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[1610,12,1980,2017,2018],"tags":[1236,1864,2046,2045,1892,2043,1965,1872,2044],"class_list":["post-13231","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai","category-announcements","category-azure-cosmos-db","category-azure-cosmos-db-tools","category-mcp","tag-ai","tag-azurecosmosdb","tag-datamodeler","tag-devtools","tag-generative-ai","tag-hosted","tag-mcp","tag-nosql","tag-shell"],"acf":[],"blog_post_summary":"<p>Building applications with Azure Cosmos DB involves more than writing queries. Developers need to make data-modeling decisions, understand how their workloads behave, and connect application data to emerging AI experiences. They also need the flexibility to work in their preferred environment, whether that is an editor, a terminal, or an AI client. At Microsoft Ignite, [&hellip;]<\/p>\n","_links":{"self":[{"href":"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-json\/wp\/v2\/posts\/13231","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-json\/wp\/v2\/users\/80443"}],"replies":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-json\/wp\/v2\/comments?post=13231"}],"version-history":[{"count":2,"href":"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-json\/wp\/v2\/posts\/13231\/revisions"}],"predecessor-version":[{"id":13273,"href":"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-json\/wp\/v2\/posts\/13231\/revisions\/13273"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-json\/wp\/v2\/media\/13236"}],"wp:attachment":[{"href":"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-json\/wp\/v2\/media?parent=13231"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-json\/wp\/v2\/categories?post=13231"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/cosmosdb\/wp-json\/wp\/v2\/tags?post=13231"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}