← All news

Latest AI news · Original source attached

GitHub Copilot gets computer use on the desktop

GitHub Copilot can now read and control desktop apps on macOS and Windows in public preview, targeting legacy software with no API, CLI or MCP integration.

Reported from GitHub

Summary written from the original first-party source — the link to it is at the end of this post.

GitHub added computer use to GitHub Copilot CLI and the GitHub Copilot app, in public preview on macOS and Windows. Copilot can read accessible app content and visual context, click controls, enter and edit text, press keys, scroll, drag and navigate workflows across applications.

The target is stated plainly: legacy or GUI-only software that has no API, no CLI and no MCP integration. That is the category where an agent tool otherwise has nothing to call, so this is GitHub addressing the long tail rather than the well-served web-app case.

On controls, Copilot asks for approval before controlling an app, and you can review or reset the list of always-allowed applications. On macOS, onboarding walks through the required Accessibility and Screen Recording permissions. You enable it with /computer on in the CLI or through Settings, Computer Use, Enable Computer Use in the app, and /computer show and /computer off handle status and shutdown.

The limits: this is a public preview rather than general availability, it is macOS and Windows only, and organization-managed settings can disable it entirely. On macOS it depends on OS-level permissions that some managed environments will not grant.

The guidance in the post is to describe the outcome, name the applications involved and state any constraints, rather than scripting the steps.

(Original source: GitHub)

AI Slop Weekly

Get this story in context, every Sunday.

The monthly rundown of what the stories above add up to — what shipped, what changed, what to actually use. 3-minute read, free forever.

Subscribe free

No spam. Unsubscribe anytime.