Programming and Software Development

JetBrains Launches Experimental Plugin to Run AI Agents Within TeamCity Workflow

JetBrains has launched a new experimental plugin for the TeamCity platform that allows for the integration of coding agents like Codex and Claude Code directly into CI/CD pipelines, facilitating code review and build error analysis.

2026-11-16
2 min read
3 views
certi.news Editorial Team
JetBrains Launches Experimental Plugin to Run AI Agents Within TeamCity Workflow

Integration of AI Agents into the CI/CD Environment

JetBrains has announced the launch of a new experimental plugin for the TeamCity platform, aimed at enabling development teams to integrate AI Coding Agents directly into their Continuous Integration and Continuous Deployment (CI/CD) workflows. This plugin allows for running tools such as Codex and Claude Code as part of build steps, opening new horizons for automating tasks such as code review, analyzing the causes of build failures, and updating technical documentation.

What changes practically for developers?

This plugin goes beyond simply running commands, as it provides an integrated infrastructure for managing agents within the development environment:

  • Automatic Management: Developers do not need to manually install tools on every build agent; the plugin checks for the client's availability and installs it automatically when the build process starts.
  • Credential Security: The plugin addresses common security risks by managing service provider credentials through TeamCity connections, where they are provided to the agent only during the build period and removed immediately upon completion.
  • Consumption Tracking: The plugin provides accurate statistics on token consumption, helping teams monitor costs and detect any abnormal patterns in model usage.
  • Trajectory Analysis: The plugin records the agent's steps and interactions, providing them in HTML and JSON formats, which facilitates the debugging process and the improvement of prompts based on actual results.

Support for Model Providers

The plugin supports a wide range of Large Language Model (LLM) providers via TeamCity connections, including JetBrains Air, OpenAI, Anthropic, as well as gateways such as LiteLLM and OpenRouter, giving teams complete flexibility in choosing the models that suit their technical needs and available budget.

Important Considerations

As the plugin is still in its experimental phase, JetBrains emphasizes the need to handle sensitive data with caution; reports generated by agents may contain snippets of source code or the prompts used, so access to these files must be restricted within the TeamCity environment. Users can install the plugin via the JetBrains Marketplace and share their feedback and experiences via the TeamCity community on Slack.

News source
JetBrains Blog
Open original source ↗
c
Author

certi.news Editorial Team

In the same category

You may also like

View all news