Azure DevOps
Azure DevOps is a set of Microsoft software delivery services: Boards for planning, Repos for Git or TFVC, Pipelines for CI/CD, Test Plans and Artifacts. It runs as a hosted service or as Azure DevOps Server on your own infrastructure, and connects to GitHub Copilot and other AI assistants through an MCP server.12
Also known as Azure DevOps Services, Azure DevOps Server, Team Foundation Server (earlier on-premises versions)
- Intended for
- Development teams, Release and platform engineers, QA and test teams, Project and product managers
- Runs on
- Cloud service (Azure DevOps Services), On-premises (Azure DevOps Server), Microsoft-hosted pipeline agents, Self-hosted pipeline agents
At a glance
- What is it?
- Azure DevOps bundles five services: Azure Boards (work items, backlogs, sprints and Kanban boards), Azure Repos (unlimited private Git repositories or Team Foundation Version Control), Azure Pipelines (build, test and deploy to Azure, AWS, Google Cloud or on-premises), Azure Test Plans (manual and exploratory testing) and Azure Artifacts (NuGet, npm, Maven, Python and Universal package feeds). Dashboards, wikis, service hooks and Power BI analytics sit alongside. Azure DevOps Services is hosted by Microsoft; Azure DevOps Server offers the same core services for organizations that keep data on their own infrastructure.1
- What does it do?
- Teams plan work items in Boards, keep code in Repos with branch policies and pull requests, run CI/CD in Pipelines with approvals and deployment gates, manage test cases in Test Plans and publish packages to Artifacts feeds. Traceability links connect work items to commits, pull requests and builds. The Azure DevOps MCP Server gives AI assistants access to work items, pull requests, builds and test plans, and a GitHub Copilot integration lets a work item start Copilot to open a draft pull request in a connected GitHub repository.134
- Who needs it?
- Organizations that want one Microsoft toolchain from planning to release, teams that already use TFVC or Azure Boards, organizations that need an on-premises option (Azure DevOps Server), and teams that deploy to several clouds from one pipeline system. It also fits mixed setups where code is on GitHub but planning stays in Azure Boards.1
- What does it need?34
- An Azure DevOps organization (free to create); a Microsoft Entra tenant behind it for the remote MCP Server
- Basic licenses for users beyond the first five free users; Basic + Test Plans for test management
- For Azure DevOps Server: your own servers and maintenance
- For the GitHub Copilot work item integration: GitHub repositories, GitHub App authentication and an active Copilot subscription
- For the local MCP Server: Node.js 20 or later
- What it is not
- Azure DevOps is not GitHub, although the two integrate: the GitHub Copilot work item integration works only with GitHub repositories, not Azure Repos. Copilot code review for Azure Repos is a separate public preview billed to an Azure subscription. The remote MCP Server needs an organization backed by a Microsoft Entra tenant and does not support standalone Microsoft account organizations. Azure DevOps Server needs your own infrastructure and maintenance.345
Availability and licensing. Prices checked 26 Oct 2026 on Microsoft's pricing page (estimates, no date shown): first 5 Basic users free, then $6 per user per month; Basic + Test Plans $52 per user per month; stakeholders free; 1 free Microsoft-hosted parallel job with 1,800 minutes, extra jobs $40; 1 free self-hosted job, extra $15; Artifacts 2 GiB free. The MCP Server is free to use. Copilot code review for Azure Repos has been in public preview since 26 Aug 2026.36
The problem it solves
Software teams often assemble planning, source control, CI/CD, testing and package management from separate tools, which breaks the link between a requirement, the code that implements it and the build that ships it.
Azure DevOps keeps these in one system with shared identity, permissions and traceability links between work items, commits, pull requests, builds and test results. It also integrates with GitHub and other Git providers, so teams can mix it with code hosted elsewhere.1
How it works
An Azure DevOps organization holds projects, and each project can use the services it needs:
- Boards: work items, backlogs, sprints and queries, with Agile, Scrum, CMMI or custom processes.
- Repos: Git (the default for new projects) or TFVC, with branch policies and pull request reviews.
- Pipelines: CI/CD on Microsoft-hosted or self-hosted agents, with approvals and deployment gates.
- Test Plans and Artifacts: test management and package feeds with upstream sources.
Identity integrates with Microsoft Entra ID. Service hooks and the Visual Studio Marketplace add integrations. For AI, the remote Azure DevOps MCP Server is a hosted endpoint over streamable HTTP that works with VS Code with GitHub Copilot, Visual Studio, Microsoft Foundry, Microsoft Copilot Studio, the GitHub Copilot CLI and the GitHub Copilot app. A local server (Node.js 20 or later) covers clients such as Claude Desktop, Claude Code, Cursor and Codex.13
Diagram as a list
Agents & automation
- Azure DevOps MCP ServerGives AI assistants access to Azure DevOps data
- GitHub Copilot (work item integration, code review)Implements work items in GitHub and reviews Azure Repos pull requestsConnects to Azure Repos (Git or TFVC)
Developer tools
- Azure BoardsPlans and tracks work items
- Azure Repos (Git or TFVC)Stores code; branch policies and pull requestsConnects to Azure Pipelines
- Azure PipelinesBuilds, tests and deploys
- Azure Test PlansManual and exploratory testing linked to work itemsConnects to Azure Boards
- Azure ArtifactsPackage feeds consumed by pipelines
Security, identity & governance
- Microsoft Entra IDIdentity for users and for the remote MCP ServerConnects to Azure DevOps MCP Server
Cloud & hybrid infrastructure
- Deployment targets: Azure, AWS, Google Cloud, on-premisesWhere releases run
Capabilities
Azure Boards14
Work items, backlogs, sprint planning, Kanban boards, queries and portfolio hierarchy from epics to tasks.
Why it matters: Keeps planning linked to code, builds and tests.
Limits: Starting GitHub Copilot from a work item works only with connected GitHub repositories.
Azure Repos14
Unlimited private Git repositories or centralized TFVC, with branch policies and pull request reviews.
Why it matters: Enforces review and build rules before code reaches the main branch.
Limits: Azure Repos is not supported by the GitHub Copilot work item integration.
Azure Pipelines16
CI/CD for any language, with deployments to Azure, AWS, Google Cloud or on-premises, parallel jobs, gates and approvals.
Why it matters: One pipeline system for several clouds and platforms.
Limits: The free tier includes one Microsoft-hosted parallel job with 1,800 minutes; extra hosted jobs are listed at $40 each per month (checked 26 Oct 2026).
Azure Test Plans and Azure Artifacts16
Manual and exploratory test management with traceability, and package feeds for NuGet, npm, Maven, Python and Universal packages.
Why it matters: Covers testing and dependency management in the same system.
Limits: Test Plans needs the Basic + Test Plans license, listed at $52 per user per month; Artifacts includes 2 GiB of free storage (checked 26 Oct 2026).
Azure DevOps MCP Server3
Remote (hosted) and local MCP servers that give AI assistants access to work items, pull requests, builds, test plans and documentation.
Why it matters: Developers can query sprint status, pull requests and pipeline health from their AI assistant.
Limits: The remote server needs an Entra-backed organization; the assistant must run in agent mode; AI platform usage may cost extra.
GitHub Copilot integrations45
Start Copilot from a work item to create a branch and draft pull request in GitHub; Copilot code review for pull requests in Azure Repos.
Why it matters: Brings agent work and automated review into Azure DevOps workflows.
Limits: Copilot runs cannot be cancelled once started. Code review for Azure Repos is a public preview since 26 Aug 2026, billed to the linked Azure subscription and rolling out by region.
Practical use cases
Planning lives in Azure Boards, but code has moved to GitHub and hand-offs are slow.
- Approach
- Connect Azure Boards to GitHub with the GitHub App and start GitHub Copilot from work items to open draft pull requests.
- Role of Azure DevOps
- Azure Boards tracks progress; Copilot implements the work item in GitHub.
- Data, infrastructure and skills
- GitHub repositories, GitHub App authentication, an active Copilot subscription and Contribute access in Azure DevOps.
- Type of benefit
- Traceability from work item to code
- Caveats
- Azure Repos is not supported for this integration, and runs cannot be cancelled once started.
- First step
- Connect one project to one GitHub repository and run Copilot on a small user story.
Developers lose time switching to the web portal to check sprint status, pull requests and builds.
- Approach
- Configure the remote Azure DevOps MCP Server in VS Code or Visual Studio and query Azure DevOps from the AI assistant in agent mode.
- Role of Azure DevOps
- The MCP Server retrieves data; the assistant summarizes it.
- Data, infrastructure and skills
- An Entra-backed organization and a supported client.
- Type of benefit
- Less context switching
- Caveats
- Standard Azure DevOps pricing applies and the AI assistant may have its own costs.
- First step
- Add the remote server in VS Code and ask for your current sprint work items.
Pull requests in Azure Repos wait for a first review.
- Approach
- Enable Copilot code review for Azure Repos at organization, project or repository level, or trigger it through branch policies.
- Role of Azure DevOps
- Copilot adds review comments before human reviewers start.
- Data, infrastructure and skills
- Azure DevOps Services and a linked Azure subscription for billing.
- Type of benefit
- Faster feedback
- Caveats
- Public preview with regional rollout; charges appear as a separate meter in Azure Cost Management.
- First step
- Enable it on one repository and set a budget alert on the project tag.
Works with
Optional integration
- GitHub CopilotWork items can start GitHub Copilot; Copilot code review is in public preview for Azure Repos.
- Visual Studio Code and Visual Studio 2026Visual Studio and VS Code are supported clients for the Azure DevOps MCP Server.
- Microsoft EntraAzure DevOps integrates with Microsoft Entra ID; the remote MCP Server requires an Entra-backed organization.
- Power BIAzure DevOps Analytics supports Power BI reporting.
- Microsoft Copilot StudioCopilot Studio is a supported environment for the remote Azure DevOps MCP Server.
- Microsoft FoundryMicrosoft Foundry is a supported environment for the remote Azure DevOps MCP Server.
Optional integration for
- GitHub CopilotAzure Boards work items can start Copilot to open a draft pull request in a linked GitHub repository.
- Visual Studio Code and Visual Studio 2026Visual Studio 18.10 supports Azure DevOps MCP servers, and the VS Code web version edits code on Azure Repos.
- Azure Machine LearningAzure DevOps pipelines are a supported CI/CD tool.
Relationship labels follow Microsoft's documentation. "Alternative approaches" does not mean one is better: each profile says when it fits.
Getting started
Create a free organization
Sign up for Azure DevOps Services and create an organization.
Check: You can open the organization and create a project.
Create the first project
Choose a process template and invite teammates; the first five Basic users are free and stakeholders are free.
Check: The backlog in Azure Boards shows your first work items.
Import or create a repository
Import an existing Git repository into Azure Repos and add a branch policy that requires review on the main branch.
Check: A pull request to main cannot complete without a reviewer.
Set up the first pipeline
Create a pipeline that builds and tests on each commit.
Check: A commit triggers a pipeline run that succeeds.
Connect an AI assistant
Configure the remote Azure DevOps MCP Server in a supported client such as VS Code with GitHub Copilot.
Check: The assistant lists your assigned work items.
Official resources
Deciding whether you need it?
Three free tools apply Microsoft's own published guidance to your situation, and say when Azure DevOps or any new technology is not needed.
Sources
Each statement above links to the source it comes from. Labels say who reported it.
- What is Azure DevOps? (opens in a new tab)
- Azure DevOps product page (opens in a new tab)
- Enable AI assistance with the Azure DevOps MCP Server (opens in a new tab)
- Use GitHub Copilot with Azure Boards (opens in a new tab)
- Copilot code reviews for Azure Repos: public preview (Azure DevOps blog, 26 Aug 2026) (opens in a new tab)
- Azure DevOps Services pricing (opens in a new tab)
Thank you. Your correction was sent.
The editors check it against the sources. If you left an email address, they may reply about it.