open-computer-use vs Verdesk

open-computer-use is a free, MIT-licensed local MCP server that lets AI agents inspect and operate desktop applications through accessibility interfaces. Its nine core tools cover app discovery, accessibility-tree and screenshot capture, clicking, scrolling, dragging, typing, key presses, secondary actions, and control-value changes. The project says it runs entirely on the user’s machine and uses the accessibility layer for core app actions without taking over the physical mouse and keyboard. Setup documentation covers Codex CLI and App, Claude Code, Gemini CLI, and opencode, with manual configuration for other MCP clients. A ZCode plugin installs the skill and connects an MCP server. macOS 14 or later requires Accessibility and Screen Recording permissions. Windows and Linux use a signed-in desktop session, and Linux requires AT-SPI2. Additional window-level tools are currently available on Windows, while macOS and Linux support is in progress. Password managers are refused; app launching, focus changes, and global input injection require explicit environment-variable gates. The project warns that high-privilege automation should be used only on authorized machines.

Verdesk is a local MCP server for Windows that lets AI agents control desktop applications and browse pages through one server. It combines browser and desktop control, offering a DOM mode for reading page code and an undetectable mode that operates a real mouse and keyboard. Native OCR reads on-screen text without sending images to the agent, while capture zones restrict attention to a selected screen area. Verdesk reports changes since the previous look and can wait for a change. Its listed uses include web-portal paperwork, extracting page data, testing applications, changing Windows settings, and repeating desktop tasks. The Free plan is for personal use and includes browser modes, OCR, capture zones, change detection, and an action log and sounds. Pro adds action scripting, button captures, and remote-PC control over SSH. It uses standard MCP and supports any model, with installation references to Claude Code, Codex, and OpenCode.