ScreenPilot
@Mtehabsim
Tool that allows the AI to control your device in the same way you do, enabling automation for everything!
概要
What is ScreenPilot?
ScreenPilot is an MCP server that lets an LLM take full control of your device by providing a screen automation toolkit for controlling and interacting with graphical user interfaces. It is built for automation, education, and having fun.
How to use ScreenPilot?
Install Python 3.12, clone the repository, create and activate a virtual environment, install the required packages from requirements.txt, then configure the server in Claude AI Desktop’s MCP settings by adding the full paths to python.exe and main.py in the config file. After saving and restarting Claude, ScreenPilot’s tools become available.
Key features of ScreenPilot
- Screen capture and analysis
- Mouse control (clicking, positioning)
- Keyboard input (typing, key presses, hotkeys)
- Scrolling in different directions and to specific positions
- Element detection (check if elements exist, wait for them)
- Action sequences (perform multiple actions in order)
Use cases of ScreenPilot
- Automate repetitive desktop tasks via an LLM agent
- Create educational demos that show step‑by‑step GUI interactions
- Build fun experiments that let an AI control on‑screen elements
- Assist with UI testing by capturing and analyzing screenshots
FAQ from ScreenPilot
What does ScreenPilot do?
ScreenPilot is an MCP server that gives an LLM the ability to capture screenshots, control the mouse, type text, press keys, scroll, detect elements, and run action sequences on your device.
What are the system requirements?
You need Python 3.12 installed, a virtual environment, and the packages listed in requirements.txt. The server is designed to work with Claude AI Desktop (the README provides configuration steps for it).
What tools does ScreenPilot provide?
Six tools: Screen Capture, Mouse Control, Keyboard Actions, Scrolling, Element Detection, and Action Sequences.
How do I install ScreenPilot?
Clone the repository, create a Python 3.12 virtual environment, activate it, run pip install -r requirements.txt, then configure the MCP server in Claude AI Desktop’s settings by editing the config file with the correct paths to python.exe and main.py.
