computer-automation

Control desktop apps through visual automation.

desktop automation

What it does

Controls local desktops or remote Windows desktops through screenshot-driven natural-language automation.

When to use it

Use for desktop-native applications that cannot run in a browser or for remote Windows hosts accessed through RDP.

Give it a desktop task; it performs the requested screen interactions, verifies the result, and reports what happened.

This skill

local desktop

Write

Controls your mouse and keyboard

remote Windows desktop

Write

Controls a remote Windows desktop

@midscene/computer@1

Execute

The Midscene computer package

desktop screenshots

Read

Reads your desktop screenshots

Requires a configured model with strong visual-grounding capabilities.

Each CLI invocation runs as a fresh node process.

Reads the model API key from the environment; Midscene can also load it from a current-directory .env file.

A reachable remote Windows desktop enables RDP mode instead of local desktop control.