You can now ask Copilot for a code review from a script

GitHub says you can now request a Copilot code review through its REST and GraphQL APIs, and set the review effort on each request. The change landed in a changelog entry on October 2 , and it’s available on Copilot Pro, Pro+, Max, Business and Enterprise.

Until now, getting Copilot to review a pull request meant clicking a button, or relying on automatic review. An API means a review can be triggered by whatever else you already automate: a label, a CI result, a release branch, a chat command.

How to request one

Per the Copilot code review docs , you request a review over REST by asking for copilot-pull-request-reviewer[bot] as a reviewer. That’s the existing endpoint for requesting reviewers on a pull request:

gh api \
  --method POST \
  repos/OWNER/REPO/pulls/123/requested_reviewers \
  -f "reviewers[]=copilot-pull-request-reviewer[bot]"

The changelog says you can optionally set the effort level per request, but it doesn’t spell out the parameter name or show a call that does it. Check the REST and GraphQL reference before you build on it. I’d rather point you at the docs than guess a field name.

The default moved

Separately, the default review effort is now Balanced rather than the previous default. GitHub says the switch took effect on September 28. If you had explicitly picked Lite, that choice was kept.

The two levels are described in the docs like this. Lite is a “cost-efficient review” aimed at glaring issues such as bugs, security vulnerabilities and style inconsistencies. Balanced does deeper analysis of complex logic, security-sensitive code and cross-service changes, using a higher-reasoning model.

So if your team never touched the setting, your reviews just got more thorough, and probably slower. That’s fine for most pull requests. It’s less fine on a repo that merges fifty small dependency bumps a day.

Where the setting lives

Defaults can be set at several levels. A September 23 changelog added a personal code review page under your profile’s Copilot settings, with a Lite or Balanced default, plus toggles for automatic review on new pushes and on draft pull requests. Enterprise admins can set one default for the whole enterprise, and organizations and repositories can override it.

What to do with it

The obvious use is matching depth to risk. A rough pattern worth trying: Lite for docs, config and dependency changes, Balanced for anything touching auth, payments or shared libraries. You could wire that to path filters or labels in a workflow and have the script pick the effort.

One caveat. The changelog is thin on detail, so the first thing to test is how the API behaves when a review is already pending or when the pull request is a draft. Run it on a throwaway repo first.