When we released the local Azure DevOps MCP Server, it gave customers a way to connect Azure DevOps data with tools like Visual Studio and Visual Studio Code through GitHub Copilot Chat. The next step was to make this experience easier to get started with and to enable it for services that support only remote MCP servers.
The Remote MCP Server is a hosted version of the Azure DevOps MCP Server that uses streamable HTTP transport. It supports the same core scenarios as the local server but removes the need for additional setup and installation.
The Remote Azure DevOps MCP Server preview is available now. We are excited to see how teams use it and will continue investing in the experience as we expand support and improve the MCP Server tools.
👟 Getting Started
Getting started in simple. Depending on the tools that you are using, you only need to add the following server information to your mcp.json.
{
"servers": {
"ado-remote-mcp": {
"url": "https://mcp.dev.azure.com/{organization}",
"type": "http"
}
},
"inputs": []
}
There are additional configuration options available, and you can read more in our official documentation.
⚙️ Supported Clients and Services
The Remote MCP Server is hosted on the Azure DevOps service and uses Microsoft Entra for authentication. As a result, it follows the authentication rules and constraints that come with Entra. This also means your Azure DevOps organization must be backed by Entra. Standalone organizations, sometimes referred to as MSAs, are not supported by the Remote MCP Server.
The following clients are supported today with no additional onboarding required:
- Visual Studio with GitHub Copilot
- Visual Studio Code with GitHub Copilot
⌛ Coming Soon
Additional client tools such as GitHub Copilot CLI, Claude Desktop, Claude Code, and ChatGPT require dynamic registration of an OAuth Client ID in Entra before they can be used with the MCP server. We are working closely with the Entra team to enable this capability. For now, only Visual Studio and Visual Studio Code are supported.
Support for other services, including Azure AI Foundry, Microsoft 365 Copilot, and Copilot Studio, is not yet available. We will share a separate blog post when these services become available with the Azure DevOps MCP Server.
📌 Local MCP Server
You can continue using the local MCP Server for the time being. However, we plan to eventually archive that repository and focus our investments on the Remote MCP Server. We do not have a specific archive date yet, but it will align with the Remote MCP Server reaching general availability.
Now is a great time to try the Remote MCP Server and begin moving your workloads over to it instead of relying on the local MCP Server.
🆘 Support
During the public preview, you can submit any issues or questions through the local Azure DevOps MCP Server repository. This allows our team to quickly review and respond to any problems or feedback you may have.
When is the Remote MCP for Azure DevOps expected to reach General Availability (GA)?
Currently planning for early-to-mid July
I’m really looking forward to the integration with Copilot Studio—does anyone know anything about it?
This will make things a lot easier for us in the future
There is an onboarding process for copilot studio that we are currently going thru. Should have something available for preview in the next 2 weeks.
The remote mcp is working really nice inside of VScode. Any updates on the Copilot Studio integration?
@Hayden
The Copilot Studio integration is admittedly a bit messy today. We are currently building a first party app that will make this much easier. In the meantime, we were able to create a connector for our MCP Server, but the onboarding experience still has quite a bit of friction.
That said, would it be helpful if we put together another blog post that walks through the setup steps in detail?
Hey Dan, hope all is well. Any update on the timeline re: onboarding process for copilot studio? Would love to use it there.
Been testing through GitHub Copilot in the meantime and it’s amazing.
Great to see the evolution from the local MCP Server to a fully hosted Remote MCP Server this significantly reduces setup complexity and improves accessibility.
This kind of integration is especially valuable for organizations dealing with large-scale data governance and DevOps workflows. In my experience working with platforms like Solix, managing data lifecycle, compliance, and lineage across systems becomes much more effective when such integrations are available.
Looking forward to seeing how this evolves further with broader client and AI agent support definitely a strong step toward more intelligent, connected DevOps ecosystems
Will we be able to use the Entra application API Permission “Azure DevOps – user_impersonation” to authenticate users to this remote MCP server? This is how we authenticate users to the API, so I’m curious/hopeful that the MCP server is the same or there will be a new API permission for MCP.