JetBrains Air

How to use JetBrains Air?

To use JetBrains Air effectively, start by downloading the macOS application and opening your local or Git-cloned project workspace. You define a specific coding task by referencing files, commits, or symbols, which provides the necessary context for the AI. Next, select an AI agent like Claude or Codex and choose an isolated execution environment, such as a Docker container or Git worktree, to safely run the task. While the agents work asynchronously, you can monitor their progress, queue follow-up prompts, and intelligently review the generated code before merging the changes into your main branch.

JetBrains Air's Core Features

• Code-aware task definitions that natively reference symbols, classes, methods, and commits.

• Concurrent execution capabilities for running multiple AI agents without project conflicts.

• Isolated execution environments utilizing Docker containers and dedicated Git worktrees.

• Built-in support for popular AI models including Claude Agent, Codex, Gemini CLI, and Junie.

• Context-aware code review interface tailored specifically for evaluating AI suggestions.

• Agent-agnostic architecture allowing seamless switching between different LLMs mid-project.

• Project-scoped workspaces for managing concurrent tasks, session history, and Git states.

• Support for Bring Your Own Key (BYOK) and integration with JetBrains AI Pro subscriptions.

JetBrains Air's Use Cases

• #1

  Orchestrating multiple AI agents to work on different codebase tasks simultaneously.

• #2

  Isolating AI coding tasks in Docker containers or Git worktrees to prevent conflicts.

• #3

  Comparing code output quality between Claude, Codex, Gemini, and Junie.

• #4

  Delegating complex or time-consuming refactoring tasks to specialized AI agents.

• #5

  Performing context-aware code reviews on agent-generated pull requests.