Azure DevOps

You can query Azure DevOps data from a custom MCP server powered by NestJS. This server exposes projects, work items and ticket statistics to any MCP-compatible AI agent, enabling real-time insights and automation across your Azure DevOps work streams.

How to use

Connect to the MCP server using either the STDIO transport for local testing or the HTTP transport for remote access. The server exposes tools you can call through your MCP client to list projects, count tickets, fetch work item details, and search work items. Use the actions to build AI-assisted workflows like sprint planning, ticket prioritization, and automated status reporting.

How to install

Prerequisites: Node.js and npm installed on your machine. You may also want to install an MCP Inspector if you plan to test STDIO mode locally.

# Install dependencies and build the project
git clone https://github.com/saktheeswar/Azure_Devops_MCP_Server.git
cd azure-devops-mcp
npm install
npm run build

Additional content

Configuration and runtime details are below. You can run the server in STDIO mode for local testing or in HTTP mode for remote access. The server requires environment variables to connect to Azure DevOps data.

BASE_URL=url
AUSER_NAME=your-personal-access-token
ADO_PAT=your-default-project

Run in STDIO mode to connect via MCP Inspector and STDIO transport. This starts the server so the inspector can attach to it.

node dist/main.js

Run in HTTP mode to expose an MCP-over-HTTP endpoint. The server will listen on port 3000 by default.

npm run start:prod

Environment and tooling

Environment variables you should prepare in a .env file are shown above. The project uses TypeScript and relies on npm and npx for package management and running MCP-related tooling.

Tools exposed by the server

The MCP server provides the following tools to interact with Azure DevOps data.

Troubleshooting and notes

If you encounter connectivity issues, ensure the HTTP endpoint is reachable and that the base Azure DevOps URLs and tokens are correctly configured in your environment. For local testing, verify that the STDIO path is active and that the MCP Inspector can attach to the running process.

Examples of usage

You can request a list of all Azure DevOps projects, a summary count of tickets, or details for a specific work item through MCP client calls to the exposed tools.

Available tools

list_projects

Returns all Azure DevOps projects

count_tickets

Returns summary of created/active/resolved tickets

get_workitem

Fetch details of a single work item

search_workitems

Query work items based on filters