Dan Hellem introduces the public preview of the Azure DevOps MCP Server. This new local service lets GitHub Copilot interact with Azure DevOps, providing AI-driven support tailored to your projects while keeping your data secure within your environment.

Azure DevOps MCP Server, Public Preview

Author: Dan Hellem

A few weeks ago at Microsoft BUILD, the upcoming Azure DevOps MCP Server was announced:

👉 Azure DevOps with GitHub Repositories – Your path to Agentic AI

Today, the local Azure DevOps MCP Server is now available in public preview.

What Does This Mean?

The Azure DevOps MCP Server enables GitHub Copilot (in both Visual Studio and Visual Studio Code) to access and interact with your Azure DevOps environment, including:

  • Work items
  • Pull requests
  • Test plans
  • Builds
  • Releases
  • Wiki pages

What is an MCP Server?

A local MCP Server (Model Context Provider) acts as a bridge between your AI assistant (e.g., GitHub Copilot) and your Azure DevOps organization. Its primary role is to inject rich, real-time project context—like work items, PRs, test plans—into LLM prompts. This allows the AI to provide better, more targeted, and context-sensitive support tailored to your Azure DevOps project.

Unlike cloud-only solutions, the MCP Server runs within your local network or development environment. It was designed to securely access your private project data without that data leaving your control.

Example Usage

The how-to guide offers multiple examples and prompts. One example:

  • Take a user story
  • Generate detailed test cases from the work item description
  • Link those test cases directly back to the original story

If you want examples for other scenarios, suggestions and requests are welcome via the repository’s issues page.

Getting Started

Initial setup is straightforward, though a few steps are required. It is advised to follow the full setup instructions for best results.

Quick overview:

  1. Visit the Azure DevOps MCP Server repository
  2. Sign in using the Azure CLI: az login
  3. Copy and paste the provided configuration into your local .vscode\mcp.json file
  4. Start the MCP Server and begin integrating with your Azure DevOps tools

Feedback

The Azure DevOps team is actively collecting feedback during this public preview stage. Any issues, ideas, or suggestions can be submitted or discussed via Issues in the repo.

For further resources, guides, and updates, refer to the Azure DevOps MCP Server repository.

This post appeared first on “Microsoft DevBlog”. Read the entire article here