JetBrains Air

One workspace to run, review and govern coding agents

Advanced
Screenshot of JetBrains Air, One workspace to run, review and govern coding agents

What is JetBrains Air?

JetBrains Air is a hub for coding agents: connect Claude Agent, Codex, Junie, Copilot, OpenCode or any ACP agent, run sessions across IDEs, browser and CLI, review diffs in context, and govern team spend and permissions.

JetBrains Air is a control layer for agent-driven software development. Instead of locking you into a single assistant, it acts as a hub that connects the agents already installed on your machine (Claude Agent, Codex, Junie, Copilot and OpenCode are named on the site) plus any agent that speaks ACP, the Agent Client Protocol. The pitch is consolidation: one place to launch sessions, watch what each agent changes, and decide what actually lands in the repository. The product ships across several surfaces. Inside JetBrains IDEs, an Air plugin lets developers keep several projects open at once, each with its own session list, and surfaces unread updates, changed files and outgoing commits across all of them. Diffs open straight from an agent session, and you can validate edits with the IDE's own tooling or leave line level comments for the agent to act on. A browser surface (air.jetbrains.cloud) runs agents in the cloud so work is not tied to one laptop or one teammate, and CLI based agents such as Claude Code and Codex are routed in through Air Gateway. Teams get shared projects in dedicated cloud environments, with VM sizing, internet access rules, secrets management and centrally configured MCP servers. Repeatable chores such as code review, release notes and issue triage can be handed to scheduled automations. Organizations get governance: per team or per person permissions over models and agents, spend limits, bring your own keys from Amazon Bedrock, OpenAI, Anthropic or another provider, and analytics on which tools are used and where AI contributes. Air Context adds semantic indexing so agents begin with an understanding of the repository rather than scanning it cold, which the site frames as faster runs at lower cost on large codebases. The fit is clearest for professional engineering teams already juggling multiple coding agents and multiple tabs. Solo developers who use one assistant inside one IDE will find the orchestration layer heavier than their workflow requires.

How do you use JetBrains Air?

  1. 1Confirm Air supports your agents
    Read the announcement and product pages to check that your current agents (for example Claude Agent, Codex or Junie) are supported, and whether any you rely on need ACP compatibility.
    JetBrains Air — Confirm Air supports your agents
  2. 2Install the Air plugin in your IDE
    Add the JetBrains Air plugin from the marketplace. Once installed, Air detects compatible agents already present on the machine and lets you connect them with a click or pull new ones from the agent registry.
    JetBrains Air — Install the Air plugin in your IDE
  3. 3Pick how you pay for AI
    Choose JetBrains-managed access with credits billed at public API rates, or bring your own subscription, API keys or custom base URL. Subscriptions can be swapped between, even mid-session.
    JetBrains Air — Pick how you pay for AI
  4. 4Route terminal agents through Gateway
    If your workflow lives in the terminal, connect CLI agents such as Claude Code and Codex through Air Gateway so their runs appear in the same workspace as IDE sessions.
    JetBrains Air — Route terminal agents through Gateway
  5. 5Add semantic context indexing
    Enable Air Context so agents start from semantic understanding of the repository rather than scanning files blindly, which the site positions as cheaper and faster on large codebases.
    JetBrains Air — Add semantic context indexing
  6. 6Set organizational guardrails
    For company rollouts, request a demo of Air Governance to set model and agent permissions, configure spend limits, add provider keys once and review usage analytics.

Pros and cons

Pros

  • Connects the agents a team already pays for instead of forcing a single vendor assistantAI
  • Diff review happens inside the session, with line comments the agent can act onAI
  • Parallel project tracking shows unread updates, changed files and outgoing commits at a glanceAI
  • Governance layer covers permissions, spend limits, bring your own keys and usage analyticsAI
  • Cloud runs and shared projects let teammates pick up agent work from another machineAI

Cons

  • Cloud runs are described as available to some customers only, with gradual rollout over coming monthsAI
  • Shifting to Air means adopting JetBrains tooling and its account requirements, which is a heavy commitmentAI
  • The value comes from orchestrating several agents, so single-agent users pay complexity for little gainAI
  • Automation and shared environment features sit behind an early access program, so not everyone can evaluate themAI
  • Cost tracking depends on configuring credits and provider keys correctly, an ongoing admin taskAI

How much does JetBrains Air cost?

Free trial

Free AI access is bundled through the Junie agent on the Lite tier, fair use applying and a JetBrains Account required. The page notes Air Teams is in early access rather than a trial.

Learn more

Pricing

The homepage states that Air comes with free AI: the Junie agent is included on the Lite tier, subject to fair use and a JetBrains Account. Teams and organizations can bring their own subscription, API keys or base URL, or use JetBrains AI credits billed at public API rates. No numeric prices are published on the page.

Learn more

Integrations

Works with Claude Agent, Codex, Junie, Copilot and OpenCode, plus any agent connectable via ACP. Provider keys from Amazon Bedrock, OpenAI, Anthropic or other vendors are supported, alongside MCP servers.

Learn more

Features

Multi-agent hub with ACP support, parallel project and session tracking, in-context diff review with line comments, cloud runs and shared team projects, scheduled automations, managed environments with VM sizing and secrets, centralized MCP servers, Air Context semantic indexing, Air Gateway for CLI agents, and governance with permissions, spend limits and analytics.

Learn more

Frequently asked questions about JetBrains Air

  • How much does JetBrains Air cost?
    The homepage states that Air comes with free AI: the Junie agent is included on the Lite tier, subject to fair use and a JetBrains Account. Teams and organizations can bring their own subscription, API keys or base URL, or use JetBrains AI credits billed at public API rates. No numeric prices are published on the page.
  • Does JetBrains Air offer a free trial?
    Free AI access is bundled through the Junie agent on the Lite tier, fair use applying and a JetBrains Account required. The page notes Air Teams is in early access rather than a trial.
  • How do you use JetBrains Air?
    The walkthrough on this page covers 6 steps: 1. Confirm Air supports your agents 2. Install the Air plugin in your IDE 3. Pick how you pay for AI 4. Route terminal agents through Gateway 5. Add semantic context indexing 6. Set organizational guardrails.
  • What platforms does JetBrains Air support?
    JetBrains Air is available on Web App, Windows, MacOS and Linux.
  • What does JetBrains Air integrate with?
    Works with Claude Agent, Codex, Junie, Copilot and OpenCode, plus any agent connectable via ACP. Provider keys from Amazon Bedrock, OpenAI, Anthropic or other vendors are supported, alongside MCP servers.
  • What are the limitations of JetBrains Air?
    Cloud runs are described as available to some customers only, with gradual rollout over coming months. Shifting to Air means adopting JetBrains tooling and its account requirements, which is a heavy commitment. The value comes from orchestrating several agents, so single-agent users pay complexity for little gain.

Status

StatusActive
Views0
Outbound clicks0
Added10/22/2026

Platforms

Web AppWindowsMacOSLinux

Pricing

FreemiumSubscription