GitHub opens Copilot code reviews to external workflow requests
GitHub now lets developers request a Copilot code review through REST and GraphQL APIs, bringing reviews into scripts and internal tools. Each request can carry a review effort setting. The release arrives alongside a reminder that Balanced became the default on September 28, while teams that explicitly chose Lite retain that preference across the existing configuration hierarchy.
Artificial Intelligence··Midday
Reviews can start from a team’s own tools
GitHub made API requests for Copilot code review generally available on October 2. Copilot is GitHub’s AI coding assistant; its review capability examines proposed code changes. Developers can now initiate a review through REST or GraphQL, two interfaces for software-to-software requests, and optionally select an effort level for that request. Scripts, existing workflows and internal tools can therefore trigger the review without starting it manually in the GitHub interface.[1], [2]
Balanced became the default before the API launch
The announcement also restates an earlier change: Default now selects Balanced for new and existing repositories and organizations. That transition took effect on September 28, following an August 28 notice. It is separate from the October 2 API release. Teams that had explicitly selected Lite kept their choice, so the default change did not reset every existing preference.[1]
Request settings sit within existing account controls
Effort settings can be configured at enterprise, organization, repository and personal-account levels. A lower level can override the setting above it. Per-request selection adds another way to specify how a review runs. The new API access is available on Copilot’s individual and organizational paid plans. GitHub’s announcement supplies no comparative measurements of review accuracy, defects found or human time saved.[1]
