JetBrains Air

What is JetBrains Air?

JetBrains Air is an Agentic Development Environment (ADE) built for modern, agent-driven software development. Developed by JetBrains, it lets developers delegate complex coding tasks to leading AI agents — including Claude Agent, OpenAI Codex, Gemini CLI, and Junie — while each agent operates in full isolation to avoid conflicts. Rather than replacing your existing IDE, Air focuses on orchestrating agents, reviewing their output, and integrating results back into your codebase. It is currently available for macOS and free during its preview period.

Key Features

• Parallel Agent Execution

  Run multiple AI agents simultaneously on separate tasks, each isolated in its own environment via Docker containers or Git worktrees to prevent conflicts.

• Multi-Agent Model Support

  Supports Claude Agent, OpenAI Codex, Gemini CLI, and Junie — each connectable via their respective accounts or API billing, with model selection per task.

• Rich Context Referencing

  Attach files, folders, Git branches, commits, symbols, terminal output, or uploaded images directly to task prompts to give agents precise, relevant context.

• IDE-Like Code Review

  Review agent-generated changes in unified or side-by-side diff views, leave inline comments on specific lines, and request targeted fixes before committing.

• MCP Server Integration

  Connect to Model Context Protocol servers to extend agent capabilities with external tools and data sources beyond the local codebase.

• Flexible Permission Controls

  Configure how much autonomy agents have — from asking permission on every action to full auto-access — giving developers granular oversight over agent behavior.

Use Cases

• Parallel Feature Development : Teams can run separate agents on independent features or bug fixes simultaneously, dramatically reducing the time spent on sequential development tasks.
• Automated Bug Fixing : Delegate bug investigation and patching to an agent with relevant commit history and code context, then review the proposed fix before merging.
• Test Generation : Spin up a dedicated agent to write unit or integration tests for existing code while another agent continues working on new features.
• Codebase Refactoring : Use isolated Docker or Git worktree environments to let agents safely refactor large sections of code without affecting the main working branch.
• Web App Verification : Instantly preview agent-generated web application changes in the built-in browser before committing, catching visual or functional regressions early.

FAQs