GitHub Copilot now runs in JetBrains Air, and its cloud tasks use a Copilot plan instead of JetBrains credits

GitHub Copilot is now available as a coding agent in JetBrains Air, GitHub announced on November 3, 2026. Developers can hand it tasks in a JetBrains development environment (IDE) or in the cloud. Cloud tasks use a Copilot plan rather than JetBrains AI credits, so a team that already pays for Copilot does not spend those credits on this work.

How it works in Air

A coding agent takes a task, works through the steps, and returns changes for a person to review. In Air, a developer selects GitHub Copilot in the agent picker and connects a Copilot account. They can then follow several agent sessions at once and review the changes next to the code, with project context and IDE tools available to the agent.

The connection uses the Agent Client Protocol (ACP), which links coding agents to development tools. Air does not replace the GitHub Copilot plugin: GitHub says developers can use the two side by side and choose the workflow that suits them.

What cloud tasks require

Cloud tasks have stricter requirements than local ones. They need a JetBrains account in a business organization with Air Teams enabled, a Copilot subscription, and an Air administrator who has enabled Copilot. They start either from the browser or from the Air plugin in the IDE.

Each cloud task runs in a Docker container, an isolated software environment, on a fresh virtual machine. All traffic passes through a proxy server, and no state carries over between tasks. Copilot can build, test, and install packages there without changing the developer's own machine.

GitHub lists four limits on cloud tasks. It does not say how much of a Copilot plan a single task uses.

• Air may check a subscription or credits only when the cloud environment starts. If they are unavailable, the task fails and the session shows an error.
• A developer cannot choose whether a given task uses a personal Copilot plan or an organization's plan.
• Air's governance and analytics features do not include Copilot usage.
• Air administrators can disable Copilot through a governance policy, but they cannot set Copilot spending limits in Air.

Usage reports now separate own-key requests

GitHub also changed the Copilot usage metrics API on November 3, 2026. Enterprise and organization reports now separate bring-your-own-key (BYOK) usage, meaning requests sent to a company's or developer's own model provider or to a local model, from models hosted by Copilot. These requests used to blend into Copilot-hosted totals, because local and custom models often share names with Copilot models.

Existing totals do not change, and report days before this release do not include the new field for custom models. Local BYOK figures rely on client telemetry, so they undercount when telemetry is turned off.