jetbrains air is one system for building software with agents

Works with the agents you already use Claude Agent
and any you can connect via ACP.

试用 Air

阅读更多内容

Choose where you work:

JetBrains IDEWebCLI移动端

air.jetbrains.cloud

Air Project

Core project behind the Air ADE — dev and analytics repos

2members2 active this week

2environmentsJCP Air Staging, JCP Air Prod

0members exceeded quotaRenews on Aug 1

300/1500credits usedRenews on Aug 1

Automations6

AutomationTriggerRunsLast RunAI Credits Spent

Agents.md updaterMerge to main14 min ago1.18

Auto-rebaserBranch behind main24 min ago6.24

Code-reviewerPR created15 min ago18.46

Documentation updateMerge to main14 min ago1.90

Issue triageIssue created126 min ago2.84

Lint checkerCommit pushed35 min ago4.52

$

9:41

Add a dark mode toggle to the site header

On it — I’ll add a dark-mode toggle to the site header and persist the choice so it survives reloads. Scanning the project: the header lives in components/Header.tsx and there’s no theme provider yet, so I’ll add a small useTheme hook alongside it.

TScomponents/Header.tsx+26−2

Done. The toggle is wired to a small useTheme hook backed by React context, and the choice persists to localStorage so the site reopens in the last-used mode. The button swaps a sun/moon icon, exposes aria-pressed, and is reachable by keyboard. I also moved the header’s hard-coded colors to CSS variables.

0.58 credits · 4s

Default to the system color scheme on first visit

Good call — I’ll read the system preference on first load and only override it once the user picks a theme, so returning visitors keep their choice.

TSlib/theme.tsx+9−1

Done. First-time visitors now inherit their OS light/dark setting, and the toggle still wins and persists after that. I also added a listener so the site follows OS changes until a manual choice is made.

0.31 credits · 3s

Follow-up on this task…

即将推出

面向开发者

run any agent and review every change with air in ides

Track parallel work

Keep several projects open, each with its own sessions, and track unread updates, changed files, and outgoing commits in all of them.

Easily connect agents

Air finds compatible agents on your machine and lets you connect them with a click. You can also add new ones from the agent registry.

Review every change in context

Open diffs directly from an agent session, validate changes with your IDE tools, and leave comments on specific lines for the agent to act on.

灵活的 AI 访问

Bring your existing provider subscription or API keys, or even base URL. You can also use JetBrains AI with credits billed at public API rates. Switch between subscriptions – even mid-session.

在您的 IDE 中试用 Explore Air in IDEs

团队

bring your team and agents together with Air Teams

Run shared projects

Teams collaborate and run agent tasks in dedicated cloud environments. Invite others and manage access from one place.

Automate busywork

Delegate repeatable tasks to coding agents. Set up automations for code review, release notes, fixes, and more.

Control project environments

Dedicated cloud environments for every project. Set the VM size, control internet access, and securely manage variables and secrets.

Centralize your setup

Connect MCP servers once so your team and their agents can securely use the same external tools.

Get early access to Air Teams

组织使用

adopt ai at scale
with Air Governance

Set rules and permissions

Give developers access to the models and agents they need. Set permissions for the whole organization or tailor them to specific teams or individuals.

Control the cost

Direct AI spend where it adds the most value, with flexible limits for teams and individual contributors.

Use any vendor

Bring your own keys from Amazon Bedrock, OpenAI, Anthropic, or any other provider. Add your key once to give all your developers access.

Optimize with analytics

Track which tools your teams use and where AI contributes the most.

Explore Air Governance 请求演示

give agents full and precise context

Reduce exploration overhead

Agents start with an understanding of your project instead of scanning the entire codebase. Air Context’s semantic indexing speeds up execution and lowers the cost of working in production-scale repositories.

Learn more about Air Context