Building Your First Pipeline in TeamCity Cloud

TeamCity Cloud, a CI/CD service from JetBrains, is also available on-premises for teams preferring self-hosted options.

The platform is also available as an on-premises option for organizations that prefer self-hosted infrastructure.

The walkthrough starts with a free 14-day trial and connects a GitHub repository to configure jobs that build and test code.

The final pipeline will have three separate jobs—one to check code, another to run tests, and a third to package the application.

Spring Pet Clinic includes an established build process and automated tests, allowing you to focus on running builds in TeamCity rather than setting up from scratch. It comes with an established build process and automated tests already in place, allowing the walkthrough to focus on running builds in TeamCity rather than setting up infrastructure from scratch.

Pipelines can be created visually through the interface or defined in YAML for version control and reproducibility.

TeamCity automatically checks out code and runs Gradle build steps while displaying a live build log showing what happens as the job runs.

A performance monitor displays CPU, disk, and memory usage on the build agent, helping teams choose appropriately sized agents based on actual resource needs.

Test results are reported as they become available during the build rather than only at the end.

TeamCity reports test results as they become available, allowing developers to see which tests have passed or failed while the job is still running. Test results are inspectable directly under a tests tab without needing to search through the entire build log.

A second job—the 'check code' job—runs the checks configured in the project. Because code checks do not depend on test results, this job and the test job can run in parallel.

The Gradle tasks for code check include 'clean check style main' and 'check style test', running independently from the test execution.

The generated jar file and checksum file are published as build artifacts by defining them under output files, providing a downloadable result from the pipeline.

TeamCity Cloud automatically spins up multiple build agents as needed, allowing tests and code checks to run on separate agents at the same time.

The packaging job waits for both the test and code check jobs to finish successfully before proceeding, ensuring dependencies are respected.

After a few pipeline runs, test profiling tracks how execution time has changed across builds, while individual test history shows whether each test passed or failed in each run.

Beyond running separate jobs in parallel, TeamCity's parallel tests feature splits Gradle, Maven, or .NET test suites into batches and distributes them across multiple available build agents.

This optimization is especially useful when test suites grow large and take longer to complete, as it maintains faster feedback to developers.

Build steps can run inside Docker containers, with an image name and tag specified or a custom image built from a Dockerfile.

This allows teams to define the specific tools and versions a build step needs, maintaining a consistent environment across runs even though TeamCity Cloud agents come with a wide selection of commonly used tools.

From the terminal, teams can inspect test results and build logs; a full list of available commands is accessible via the teamcity -h flag.

AI coding agents can use the TeamCity CLI to interact with builds and pipelines. Agent skills provide guidance for working with TeamCity, including help migrating existing workflows from other CI/CD tools.

GitHub Actions and Bamboo workflows are currently supported for migration, with additional options coming in the future.

TeamCity includes a built-in Model Context Protocol (MCP) server, enabling compatible AI tools to connect, inspect builds, and help investigate failures.

A read-only access token created in TeamCity can be used to connect an IDE's AI assistant to TeamCity's MCP server, allowing questions about builds to be asked directly in the AI chat interface.

The TeamCity IDE plugin for JetBrains IDEs displays pipeline status and test results directly within the editor, allowing developers to monitor builds while working on code without switching applications.