TeamCity MCP Server: Let AI Monitor Builds and Deployments

Why Give AI Access to TeamCity?

TeamCity is the CI/CD server behind many enterprise Java, .NET, and polyglot teams. When you need to check if a build passed, review test results, or verify a deployment completed, you switch between your IDE and the TeamCity web UI. During release cycles, those interruptions stack up.

An MCP server for TeamCity lets your AI agent query builds, check pipeline status, review test failures, and inspect deployment history directly from Claude, Cursor, or Windsurf.

What Becomes Possible

Once your AI agent has TeamCity access:

• "Did the latest build on main pass?"
• "What tests failed in the last CI run?"
• "Show me the deployment history for the payments service"
• "Is there a build running right now?"
• "What was the build duration trend this week?"
• "Which agent ran the last production deploy?"

How It Works

1. Go to DataFaucet and enter your TeamCity URL (e.g. https://teamcity.company.com)
2. Log into your TeamCity instance
3. Browse builds, configurations, and agents as you normally would
4. Each action becomes an MCP tool your AI can call
5. Deploy and connect to Claude, Cursor, or Windsurf

Works with TeamCity Cloud and self-hosted. The AI agent gets whatever access your session provides.

Security Best Practices

• Read-only tokens — create API tokens with only view-build-configuration and view-project permissions
• Project-scoped — restrict tokens to specific projects, not server-wide
• No trigger permissions — prevent AI from starting or canceling builds without approval
• IP restrictions — limit token usage to known CI/CD infrastructure
• Audit trail — TeamCity logs all API access with token attribution

Use Cases by Role

Backend developers: Build status mid-coding. "Did my last push pass CI?" without opening TeamCity.

Release managers: Deployment tracking. "What builds deployed to production today?" for release coordination.

QA engineers: Test failure analysis. "Which tests failed in the last nightly run?" for triage.

Platform engineers: Infrastructure monitoring. "Are all build agents online and idle?" for capacity checks.

TeamCity UI vs AI Agent

Getting Started

Create a read-only authentication token in TeamCity:

1. Go to Profile → Access Tokens → Create Token
2. Set scope to specific projects (not server-wide)
3. Grant only "View project and all parent projects" permission
4. Set expiration date

Then point DataFaucet at your TeamCity instance, authenticate, and your AI agent can query builds and pipelines as MCP tools.