Desktop Control

MCP server for desktop automation via @nut-tree-fork/nut-js

Universal agent-to-software runtime. Any AI agent controls any software through a single protocol.

A browser for LLMs, not humans. Real WebKit over a line protocol built for agents. Try it with npx vibesurfer.

Give AI agents eyes and hands on macOS — screenshot, OCR, click, type through the Model Context Protocol

Talkie CLI for macOS — voice memos, dictations, screenshots, screen clips, search, and agent workflows

Universal macOS app automation for Claude — 30-100x faster than screenshot-based computer-use. Uses Accessibility API, AppleScript, and CGEvent.

MCP server for snap-x: lets any agent (Claude Desktop, Cursor, Windsurf, …) render, check and preview platform-ready branded graphics, with built-in platform formats, safe-zone guides and a design guide.

🪟 Darbot Windows MCP - Professional Windows desktop automation server for AI agents via Model Context Protocol (MCP). Control Windows like a human with 15 powerful tools. Includes Python dependencies, UV, and VS Code configuration.

Screenshot any URL and see it, or mint a social card — hosted, SSRF-safe, cached. No Chrome to run.

MCP server for desktop automation - mouse, keyboard, screenshot, window management

Evaluates whether an AI agent is safe to operate internal web apps via an MCP run_suite tool.

MCP server for accurate screenshot coordinates with small, fast vision models.

MCP Server for terminal computer use — control interactive TUI programs (vim, htop, lazygit, REPLs, installers) via MCP protocol

MCP server providing computer use capabilities through an isolated sandbox environment

MCP server that provides computer control capabilities, like mouse, keyboard, OCR, etc. using PyAutoGUI, RapidOCR, ONNXRuntime. Similar to 'computer-use' by Anthropic. With Zero External Dependencies.

Hyprland/Wayland computer-use MCP: grim screenshots + uinput click/type + hyprctl targeting

Drive your Mac focus-free over MCP: OS APIs, AppleScript, CDP, and Accessibility.

Design-first MCP for coding agents that improves websites with screenshot-grounded iteration.

Windows desktop-control MCP server: screenshot, window mgmt, mouse/keyboard input, recording.

Desktop automation MCP — screenshot (calibrated), mouse/keyboard via xdotool, window inspection