Skip to main content
Back to timeline

JetBrains brings coding agents into the IDE with Air: one place to track, review and send back changes, shipping in 2026.3

Synopsis

JetBrains shipped Air in its 2026.3 IDEs, bringing coding agents into the developer's existing environment with one place to track session progress, review changes and send feedback. JetBrains frames the problem as a shift in bottleneck: as agents write more code, review becomes the constraint, and it cites a report finding that 49% of developers now spend four or more hours a week reviewing AI output and correcting prompts. Through the IDE's MCP server and built-in skills, agents can run inspections, build and run configurations, query databases and use the IDE's own refactorings, and in IntelliJ IDEA and CLion they can also drive the debugger.

AI-generated editorial illustration: JetBrains IDEs Transformed for the Agentic Era of Development – With Air

Interpretation

JetBrains released Air in its 2026.3 IDEs, letting developers bring coding agents into their existing development environment and track session progress, review changes and send feedback in one place. JetBrains states that as agents write more code, individual productivity rises but the bottleneck moves to review, and cites the Atlassian State of Developer Experience Report 2026 finding that 49% of developers spend four or more hours a week reviewing AI output and correcting prompts. Air is available in JetBrains IDEs from version 2026.3 and works with agents such as Codex, Claude Agent, GitHub Copilot and Junie, connecting to other compatible agents through ACP.

Because Air runs inside the IDE, it brings IDE code intelligence to both the developer and the agents. Agents can use the IDE's MCP server and built-in skills to check inspections, run build and run configurations, query databases and use the IDE's own refactorings, and in IntelliJ IDEA and CLion they can also drive the debugger. When an agent finishes, the developer opens the review panel to see what the session changed, checks it with navigation, inspections, tests or the debugger, and can leave Code Notes on specific lines that go back to the session with the code context attached.

Parallel sessions each get their own local Git worktree so they do not collide, and notifications tell the developer when an agent needs input. Worktrees that agents create on their own are not yet tracked in Air, and JetBrains says it is working on that. Air can save the agent, model and instructions for a task type as a preset and start new sessions from it; local agent workflows do not require a paid JetBrains AI subscription, since an existing provider account or API key works, and Junie Lite is free with a JetBrains Account.

Perspective

The report covers JetBrains' description of Air's features: available in JetBrains IDEs from version 2026.3, supporting local agent workflows, with Air and agents running on the backend in remote development. It suits developers already using JetBrains IDEs and coding agents who are deciding whether to change their workflow.

JetBrains says worktrees that agents create on their own are not yet tracked in Air and that it is working on this, and the same usage view for Codex, Claude and Copilot subscriptions is still in development. Air does not automatically start agents or send code to an AI provider; data handling depends on the chosen agent and provider terms, and provider charges, usage limits and fair-use terms still apply.