GitHub Copilot is now available through JetBrains Air, allowing developers to delegate coding tasks, supervise AI agents and review their work from supported JetBrains development environments.
GitHub announced the integration on November 3.
The new connection uses the Agent Client Protocol, allowing Copilot to operate through the JetBrains Air interface.
Developers can run supported tasks locally or use cloud environments, depending on their account and organisational configuration.
Copilot added to Air agent selection
Developers can select GitHub Copilot from the agent picker in JetBrains Air.
The integration allows multiple agent sessions to be managed while developers review code changes using project context and development tools.
GitHub says Copilot can operate alongside its existing plugin for JetBrains development environments.
The new integration provides an additional workflow for users who want to assign tasks to coding agents and review their output.
It does not require developers to abandon the existing Copilot plugin.
Cloud tasks use GitHub subscriptions
GitHub confirmed that Copilot cloud tasks launched through JetBrains Air use the developer’s GitHub Copilot subscription.
They do not consume JetBrains AI credits.
Cloud access requires a JetBrains account belonging to an organisation with Air Teams enabled.
Users must also have an eligible Copilot subscription and administrator permission to use the integration.
GitHub says cloud tasks can be started from a browser or through the Air plugin in a supported JetBrains development environment.
The announcement does not establish that every individual Copilot subscriber automatically receives cloud access.
Cloud agents operate in isolated environments
GitHub says Copilot cloud tasks run inside Docker containers on newly created virtual machines.
Network traffic passes through a proxy.
The company states that task environments do not retain state between sessions.
Within these environments, agents can build software, run tests and install packages without directly modifying the developer’s local machine.
The arrangement separates cloud task execution from the local development environment.
Developers still need to review generated code and understand the permissions granted to coding agents.
Administrative controls have limitations
GitHub says administrators can disable Copilot through JetBrains Air governance policies.
However, Air administrators cannot configure Copilot spending limits through those controls.
The company’s announcement also states that Air governance and analytics features do not include Copilot usage.
These limitations may affect organisations that require consolidated reporting of AI coding expenditure and activity.
GitHub has not announced a date for adding Copilot usage data to JetBrains Air analytics.
Integration available following November release
Developers can begin using the integration by installing Air in a supported JetBrains environment, selecting Copilot and connecting their GitHub account.
Organisations planning cloud deployments must also configure the relevant administrative permissions.
GitHub has not published performance benchmarks comparing Copilot’s operation inside JetBrains Air with its existing development integrations.
The November 3 release establishes availability and supported workflows rather than a measured improvement in coding speed or accuracy.
GitHub Changelog, Copilot in JetBrains Air, 3 November 2026
