Skip to main content
Glama
Sign In
localScreenControl

localScreenControl

by Akicou
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Local

localScreenControl

An MCP server that lets an AI assistant see and control your local desktop.

Tool

What it does

screenshot(monitor=1, max_width=1600)

Screenshot of a monitor (1 = primary, 0 = all), downscaled to max_width

click(x, y, button="left", clicks=1)

Click at a pixel of the last screenshot (clicks=2 = double-click)

scroll(x, y, amount=-5)

Mouse wheel at a point (negative = down)

press_keys(keys)

Key combos like ctrl+tab, alt+f4, ctrl+shift+t, enter

type_text(text, paste=True)

Enter text in the focused field, via clipboard paste (default) or real keystrokes

get_clipboard() / set_clipboard(text)

Read / write clipboard text (ctrl+a, ctrl+c, get_clipboard reads a whole page)

wait_until_stable(timeout_ms=5000)

Wait until the screen stops changing (page loads, dialogs, animations)

batch(steps, screenshot_after=True)

Run many actions in one call, then return a screenshot

batch example:

[{"action": "click", "x": 700, "y": 140},
 {"action": "type", "text": "Hello"},
 {"action": "keys", "keys": "enter"},
 {"action": "settle"},
 {"action": "wait", "ms": 300}]

Works on Windows and Linux (X11). Requires Python 3.10+.

Setup

Windows (PowerShell)

git clone https://github.com/Akicou/localScreenControl
cd localScreenControl
powershell -ExecutionPolicy Bypass -File .\setup.ps1

Linux

git clone https://github.com/Akicou/localScreenControl
cd localScreenControl
./setup.sh

The script creates a .venv, installs dependencies, registers the server with Claude Code (if claude is on your PATH) and prints a JSON config snippet for other MCP clients such as Claude Desktop.

Related MCP server: screenmcp-claude-plugin

Notes

  • Coordinates are pixels of the most recent screenshot image. The server maps them back to the real screen (downscaling, DPI, monitor offset), so they always match what the AI saw.

  • Linux clipboard needs xclip or xsel (sudo apt install xclip).

  • Self-check: .venv/bin/python test_server.py (or .venv\Scripts\python test_server.py).

  • Linux on Wayland: only X11/XWayland windows can be captured/controlled. Use an X11 session for full control.

  • Key names follow pynput's Key (ctrl, shift, alt, cmd/win, tab, enter, esc, f1–f20, up, page_down, ...).

Security

This server gives the connected AI full control over your mouse and keyboard. Only connect it to clients you trust.

License

MIT

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    This server cannot be deployed

    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    Maintenance

    ActivityMaintained
    ResponsivenessNo recent community issues

    Resources

    Related MCP Connectors

    Related MCP Servers

    • servo-mcp

      d11r
      A
      license
      Not graded
      quality
      D
      maintenance
      Enables AI agents to see and control your desktop with tools for screenshots, clicks, typing, and more, all locally on macOS and Windows.
      210 npm
      MIT

    • Open Computer Use

      aibo204
      A
      license
      Not graded
      quality
      B
      maintenance
      Enables AI agents to observe and control the user's live desktop by listing apps, reading accessibility trees, and performing clicks, typing, scrolling, dragging, and other input actions with per-action approvals and local audit archives.
      1
      MIT

    • OpenMouse MCP

      osca5771666
      A
      license
      A
      quality
      B
      maintenance
      Enables agents to perceive and control a Windows desktop, a connected local Chromium browser, or a cloud Playwright browser through human-like mouse moves, clicks, drags, scrolling, and screenshots, forming an observe-locate-act computer-use loop. Includes safety guards such as random pauses, emergency stop, and approval prompts for destructive actions.
      11
      MIT