Skip to main content

Computer Split Screen MCP

A high-performance, cross-platform Model Context Protocol (MCP) server that provides reliable split-screen window management for both Windows and macOS. This server exposes 16 window manipulation tools through the MCP protocol, enabling AI assistants and other MCP clients to control desktop window layouts with precision.

🚀 Features

  • Cross-platform support: Windows and macOS with native optimizations
  • 16 window management tools: Comprehensive split-screen layouts and controls
  • High performance: Windows: 15-60ms, macOS: 50-150ms per operation
  • Reliability first: Robust fallback mechanisms and error handling
  • MCP integration: Full Model Context Protocol server support via stdio transport
  • Smart focus detection: Advanced window focus detection with platform-specific optimizations

🛠️ Installation

Prerequisites

  • Python: 3.9 or higher
  • Package Manager: uvx (recommended) or pip

Install via uvx (Recommended)

uvx install computer-split-screen-mcp

Install via pip

pip install computer-split-screen-mcp

🔧 MCP Client Configuration

Configure your MCP client with the following settings:

{
  "mcpServers": {
    "computer-split-screen": {
      "command": "uvx",
      "args": ["computer-split-screen-mcp"],
      "env": {}
    }
  }
}

Alternative Configuration (if not using uvx)

{
  "mcpServers": {
    "computer-split-screen": {
      "command": "python",
      "args": ["-m", "computer-split-screen-mcp"],
      "env": {}
    }
  }
}

🎯 Available Tools

Split-Screen Layouts

Halves (2-way splits)

  • left-half-screen - Snap current window to left half
  • right-half-screen - Snap current window to right half
  • top-half-screen - Snap current window to top half
  • bottom-half-screen - Snap current window to bottom half

Quadrants (4-way splits)

  • top-left-screen - Top-left quadrant (1/4 screen)
  • top-right-screen - Top-right quadrant (1/4 screen)
  • bottom-left-screen - Bottom-left quadrant (1/4 screen)
  • bottom-right-screen - Bottom-right quadrant (1/4 screen)

Thirds (3-way splits)

  • left-one-third-screen - Left third (1/3 screen)
  • middle-one-third-screen - Middle third (1/3 screen)
  • right-one-third-screen - Right third (1/3 screen)

Two-Thirds (2/3 splits)

  • left-two-thirds-screen - Left two-thirds (2/3 screen)
  • right-two-thirds-screen - Right two-thirds (2/3 screen)

Window Controls

  • maximize-screen - OS maximize (bordered, taskbar visible)
  • fullscreen-screen - Fullscreen mode (platform-specific behavior)
  • minimize-screen - Minimize window to taskbar/dock

⚡ Performance Characteristics

Windows Performance

  • Total time: 15-60ms per operation
  • Detection: 2-5ms (direct Win32 API)
  • Manipulation: 10-47ms (SetWindowPos with corrections)
  • Best case: 15-25ms for simple splits
  • Typical case: 20-35ms for most operations

macOS Performance

  • Total time: 50-150ms per operation
  • Detection: 20-50ms (AppleScript + Accessibility API)
  • Manipulation: 15-55ms (AXUIElement operations)
  • Best case: 50-80ms for simple splits
  • Typical case: 80-120ms for most operations

🔍 Technical Architecture

Windows Implementation

  • Direct Win32 API calls via pywin32
  • DWM integration for accurate frame bounds
  • DPI-aware positioning for high-resolution displays
  • Two-pass positioning for precise window placement

macOS Implementation

  • AppleScript focus detection for reliability
  • Accessibility API for window manipulation
  • Fallback mechanisms for edge cases
  • Screen-aware positioning respecting menu bar and Dock

Cross-Platform Features

  • Automatic platform detection
  • Conditional dependency loading
  • Unified API interface
  • Error handling and recovery

📋 Platform Dependencies

Windows

  • Required: pywin32>=306
  • Purpose: Win32 API access, DWM integration, window manipulation

macOS

  • Required: pyobjc-core>=10.1,<11, pyobjc-framework-Cocoa>=10.1,<11, pyobjc-framework-Quartz>=10.1,<11, pyobjc-framework-ApplicationServices>=10.1,<11
  • Purpose: Accessibility API, AppleScript integration, window management

🚨 Troubleshooting

Common Issues

macOS Focus Detection Problems

  • Symptom: Windows don't move or wrong windows are selected
  • Solution: Ensure Terminal has Accessibility permissions in System Preferences > Security & Privacy > Privacy > Accessibility

Windows DPI Issues

  • Symptom: Windows positioned incorrectly on high-DPI displays
  • Solution: The server automatically handles DPI awareness, but ensure pywin32>=306 is installed

MCP Client Connection Issues

  • Symptom: Functions work when called directly but fail via MCP
  • Solution: Check MCP client logs, ensure proper configuration, restart MCP client

Performance Optimization

  • First run: May be slower due to system warm-up
  • Subsequent runs: Should be consistently fast
  • Complex apps: Safari, Chrome may take longer due to window structure

🔧 Development

Project Structure

computer-split-screen/
├── src/splitscreen_mcp/
│   ├── __init__.py          # Package initialization
│   ├── __main__.py          # MCP server entry point
│   └── window_actions.py    # Core window management logic
├── pyproject.toml           # Project configuration
├── README.md               # This file
└── LICENSE                 # MIT License

Building from Source

git clone https://github.com/Beta0415/computer-split-screen-mcp.git
cd computer-split-screen-mcp
uvx install -e .

Running Tests

# Test window detection
python3 -c "from src.splitscreen_mcp.window_actions import left_half_window; left_half_window()"

# Test MCP server
uvx run computer-split-screen-mcp

Contributing

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request. For major changes, please open an issue first to discuss what you would like to change.

📊 Version History

  • v1.4.4 - Current stable release
    • Cross-platform window management
    • 16 comprehensive tools
    • High-performance implementation
    • Full MCP protocol support

🆘 Support

If you encounter any issues or have questions:

  1. Check the troubleshooting section above
  2. Review MCP client logs for errors
  3. Test functions directly to isolate issues
  4. Open an issue on the project repository

Built with ❤️ for the MCP community

Metadata

Release files for computer-split-screen-mcp 1.5.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for computer-split-screen-mcp 1.5.0
File Size Uploaded
computer_split_screen_mcp-1.5.0.tar.gz 11.1 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for computer-split-screen-mcp 1.5.0
File Interpreter ABI Platform
computer_split_screen_mcp-1.5.0-py3-none-any.whl Python 3 none any Details

Total release size: 24.1 kB

Release files / computer_split_screen_mcp-1.5.0.tar.gz

Download URL computer_split_screen_mcp-1.5.0.tar.gz
Size 11.1 kB
Tags Source
SHA-256 checksum
How to use checksums
393102737b80548c6d7d0fe5fc300f52aec608007a710d35a0c2ef8999255a3e
BLAKE2b-256 checksum
How to use checksums
fb594c7846c45b2b58d74b9e18a4b053dafd366bb5dc2935a1c9c95ce53ed668
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.1.0 CPython/3.13.3

Release files / computer_split_screen_mcp-1.5.0-py3-none-any.whl

Download URL computer_split_screen_mcp-1.5.0-py3-none-any.whl
Size 13.0 kB
Tags Python 3
SHA-256 checksum
How to use checksums
63bf97ce95d4b5bfd5936a3906e1904ec4d9cbab6d52f98c7443a9cea3624d5b
BLAKE2b-256 checksum
How to use checksums
c96180537d05e3bed33747a05a8430619d5091e33656f5f9f1252a19014289e8
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.1.0 CPython/3.13.3

Release history Release notifications | RSS feed

This release

1.5.0 This release

2 release files

1.4.5

2 release files

1.4.4

2 release files

1.4.3

2 release files

1.4.2

2 release files

1.4.1

2 release files

1.4.0

2 release files

1.3.5

2 release files

1.3.4

2 release files

1.3.3

2 release files

1.3.2

2 release files

1.3.1

2 release files

1.3.0

2 release files

1.2.0

2 release files

1.1.0

2 release files

1.0.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page