网页
综合
更新时间 2026-11-03 00:00:00
Use Your GitHub Copilot Plan in JetBrains Air – in Your IDE and in the Cloud - The JetBrains Blog
自动蜘蛛
blog.jetbrains.com
正文约 4531 字
An agentic development environment
Read this post in other languages:
GitHub and JetBrains are bringing GitHub Copilot to JetBrains Air. Air lets you hand coding tasks to agents, run them in parallel, and review their changes in one place. You can now pick GitHub Copilot as the agent in Air – locally in your JetBrains IDE, and in the cloud with Air Teams. Tasks run on your existing Copilot plan, not on JetBrains AI credits.
Air runs the official GitHub Copilot CLI over the Agent Client Protocol (ACP), the open protocol Air uses to connect agents. You get the real GitHub Copilot agent and the models it offers you, in the same task list as Claude, Codex, and Junie. With Air Teams, you can hand GitHub Copilot a cloud task and close your laptop, and it’ll keep working on its own sandboxed machine.
Try it now in your JetBrains IDE or on the web.
GitHub Copilot runs locally in Air in JetBrains IDEs through ACP. You don’t need Air Teams, but you do need a GitHub Copilot subscription.
GitHub Copilot works in your local project, and you review its changes in Air as with any other agent.
With Air Teams, GitHub Copilot can also run in the cloud. You can start a cloud task from your IDE or the web, and it keeps working after you close your laptop.
We recommend starting with a small, well-defined task – maybe something you’d normally put off: a flaky test, a dependency bump, or a small bug with a clear repro.
The task shows up in the Tasks list under your repository, both in your IDE and on the web. GitHub Copilot starts in a fresh cloud machine. When it’s done, open the task and review its changes.
Each task gets its own virtual machine. GitHub Copilot runs in a Docker container, signed in to GitHub with your account. All its traffic goes through a proxy, and nothing carries over to the next task. So it can build, test, and install packages without touching your laptop. For more details, check out our blog post How Air Cloud-Based Agent Sandboxing Works in Air.
Automations start cloud tasks for you – either on a schedule or when something happens in your repository, such as a new pull request. If you pick GitHub Copilot as the agent, each run uses your Copilot plan.
We recommend starting with a weekly dependency update as your first automation: It’s easy to review and keeps credit usage predictable. In contrast, an automation that reviews every pull request can drain your credits much faster, especially if the repository is busy. Keep in mind that Air doesn’t warn you when your credits are all used up – your runs will just start failing.
How is GitHub Copilot different from the other agents in Air?
GitHub, not Air, handles your Copilot account. When a task starts, the Copilot CLI checks your plan and credits with GitHub. Because Air doesn’t show your Copilot usage and can’t limit it, you have to keep track of it on GitHub.
Which GitHub Copilot plans can I use?
Any plan that GitHub accepts for the Copilot CLI. Air doesn’t check which specific plan you’re using.
Why did my task fail soon after it started?
Most likely, your GitHub plan has no credits left. Air checks that you have an active Copilot subscription before it starts a task, but it can’t see how many credits you have left. Check your usage on GitHub.
How do I know how many credits were spent on a task?
Check your usage on GitHub. Air doesn’t track Copilot credits for local or cloud tasks, and it can’t set a spending limit.
By submitting this form, I agree that JetBrains s.r.o. ("JetBrains") may use my name, email address, and location data to send me newsletters, including commercial communications, and to process my personal data for this purpose. I agree that JetBrains may process said data using third-party services for this purpose in accordance with the JetBrains Privacy Policy. I understand that I can revoke this consent at any time in my profile. In addition, an unsubscribe link is included in each email.
Thanks, we've got you!
Local agents share your machine. Remote ones get isolation by default. We looked into how these five setups compare and how Air sandboxes cloud agents.
“Just because you can build 100 things doesn’t mean you should.” Engineering teams share what they’re learning about building with AI agents.
Agentic workflows used to live on one developer's laptop. With Air Teams, you configure a workflow once, and the whole team can run, validate, and improve it.
Air now puts multiple projects in one window, renders Markdown files as formatted documents, and fixes IME input on Windows.
本页为82搜索生成的网页快照,内容来自站内索引;演示数据的原文链接仅作示例。