trackmcp
Back to directory
shivman

paint-ai-agent

View on GitHub

This project demonstrates an Agentic AI system that uses a Gemini-powered MCP server to interpret natural language instructions and automate Microsoft Paint using GUI simulation tools on Windows.

2 stars PythonAI & Machine Learning Updated Aug 11, 2025

Documentation

Paint Drawing Agent

A Python-based automation tool that uses Google's Gemini AI to control Microsoft Paint through natural language commands. This project enables users to draw shapes, write text, and manipulate colors in MS Paint using simple English instructions.

Features

  • Natural language control of MS Paint
  • Automated drawing of shapes (circles, rectangles, lines)
  • Text insertion with position control
  • Color selection and management
  • Window management and canvas positioning
  • Detailed logging and error handling
  • Position calibration system

Prerequisites

  • Windows 10 or later
  • Python 3.8+
  • Microsoft Paint (mspaint.exe)
  • Google Cloud API key for Gemini AI

Installation

1. Clone the repository:

bash
git clone [repository-url]
cd paint-drawing-agent

2. Install required dependencies:

bash
pip install -r requirements.txt

3. Create a `.env` file in the project root and add your Google API key:

code
GOOGLE_API_KEY=your_api_key_here

Project Structure

code
paint-drawing-agent/
├── talk2mcp.py           # Main application file
├── tools/                # Core automation tools
│   ├── __init__.py
│   └── paint_commands.py # Paint automation commands
├── calibration_profiles/ # Stored calibration data
├── LLM_LOGS/            # AI interaction logs
├── logs/                # Application logs
└── requirements.txt     # Project dependencies

Usage

1. Start the Paint Drawing Agent:

bash
python talk2mcp.py

2. Enter natural language commands, for example:

  • "Draw a red circle in the center"
  • "Write 'Hello World' in black at the top"
  • "Draw a blue rectangle on the right side"

3. To exit the program, type 'quit'

Command Examples

code
> Draw a red circle in the center
> Write 'Hello World' in black at the top
> Draw a blue rectangle at coordinates 400,300
> Draw a green line from top to bottom

Calibration

The system uses a calibration system to accurately locate Paint interface elements. To recalibrate:

1. Run the calibration script:

bash
python enhanced_calibrate.py

2. Follow the on-screen instructions to calibrate tool positions

Logging

The system maintains several types of logs:

  • Application logs: `/logs/paint_agent_[timestamp].log`
  • LLM interaction logs: `/LLM_LOGS/session_log.json`
  • Calibration logs: `/calibration_profiles/`

Error Handling

The system includes comprehensive error handling for:

  • Window management issues
  • Drawing operation failures
  • AI communication errors
  • Position calibration problems

Contributing

1. Fork the repository

2. Create a feature branch

3. Commit your changes

4. Push to the branch

5. Create a Pull Request

Troubleshooting

Common issues and solutions:

1. Paint Window Not Found

    2. Drawing Position Issues

      3. AI Communication Errors

        License

        [Your License Here]

        Acknowledgments

        • Google Gemini AI for natural language processing
        • PyAutoGUI for GUI automation
        • Win32GUI for Windows interaction

        Frequently asked questions

        What is paint-ai-agent?

        paint-ai-agent is This project demonstrates an Agentic AI system that uses a Gemini-powered MCP server to interpret natural language instructions and automate Microsoft Paint using GUI simulation tools on Windows.

        How do I install paint-ai-agent?

        Open the GitHub repository and follow its README. Most MCP servers are added to your client's MCP config, then called by your agent.

        Is paint-ai-agent open source?

        Yes — it is hosted on GitHub at https://github.com/shivman/paint-ai-agent and has 2 stars.

        Related MCP tools

        Run your own MCP server? See who uses it and what to fix.

        Measure it with TrackMCP