GitHub Copilot support in JetBrains Air is worth testing carefully because local and cloud tasks are not the same thing. JetBrains and GitHub have documented that Copilot can be used in Air for local work, and that cloud usage depends on additional account and organization setup. The safest way to evaluate it is to keep the two paths separate, use a test repository, and compare the results without assuming they will match exactly.
What GitHub announced on November 3
GitHub’s November 3, 2026 announcement covers running Copilot in isolated cloud environments through Air Teams as well as using it for local tasks in Air. The Copilot agent in Air and the separate GitHub Copilot IDE plugin have different entry points and features.
GitHub changelog: GitHub Copilot is now available in JetBrains Air
Microsoft Dev Blog: GitHub Copilot in JetBrains Air
Local task testing: keep it simple and harmless
Install Air in your JetBrains IDE, select GitHub Copilot in Air’s agent picker, and connect your Copilot account. Open a test repository and ask for a small change, such as adding a unit test for an existing function. Review the diff and run the test. Here, local describes where the task runs; it does not mean Copilot works offline.
A good local test should avoid private code, secrets, production configuration, and anything that could create risk if the generated suggestion is wrong. Use a test repo with synthetic content so you can review the output freely and repeat the task if needed.
Cloud task testing: verify the account path first
For a cloud task, use a JetBrains account in an Air Teams-enabled business organization, confirm your GitHub Copilot subscription, and have your Air administrator enable Copilot. These cloud tasks use your Copilot plan rather than JetBrains AI credits. GitHub notes that entitlement checks may occur when the cloud environment starts, so connecting an account alone does not prove cloud access.
In the IDE, choose Cloud from Air’s environment selector. In the browser, connect GitHub under Settings > VCS Providers, then create a task for the test repository and its remote branch with Copilot selected. Start from the same commit as the local test and use the same prompt so you can compare the two diffs and test results.
How to compare local and cloud results
- Use the same test repository structure for both runs.
- Choose tasks of similar scope, such as a short code explanation or a small refactor.
- Review output quality, completeness, and any differences in context handling.
- Check whether the cloud and local runs require different approvals, permissions, or setup steps.
- Do not assume identical environments or identical results.
A useful comparison notes what changed in the workflow, not just which answer looked better. For example, one environment may surface different context, prompt handling, or permission behavior. Those differences matter more than a quick winner-and-loser comparison.
A safe evaluation checklist
- Confirm you are testing in a throwaway or non-sensitive repository.
- Authenticate GitHub for the local path before starting.
- Confirm Air Teams access, Copilot subscription, and admin enablement for the cloud path.
- Use harmless prompts and avoid exposing secrets or proprietary material.
- Record the results separately for local and cloud tasks.
- Review what Copilot changed before accepting anything.
If you are documenting the test for a team, keep the notes factual: what account setup was required, what task was run, what the output looked like, and where the experience differed. That is more useful than broad claims about speed or quality that were not measured carefully.
How to interpret the result
JetBrains Air can use GitHub Copilot in both local and cloud scenarios, but the paths have different setup requirements and should be tested separately. Start with a local harmless task, then verify the account and organization requirements before trying a cloud task. Use a test repo, keep the data safe, and compare the results on their own terms.
