Chess Position Evaluator

Features

• FEN Extraction: Determines the FEN of the current board position using Zai-Kun's Chess Pieces Detection, ensuring accurate move analysis.
• Optimized Model Inference: The local ONNX model chess_detection.onnx enhances speed and eliminates external API dependencies.
• Automatic Board Flipping: Automatically flips the board if you’re playing as Black.
• Stockfish Integration: Uses the Stockfish chess engine to suggest the best move.
• Auto Move Execution: Automatically plays the best move suggested by Stockfish.
• Auto Play Moves: When enabled, the script will automatically make moves after the opponent’s move without requiring any further clicks.
• Manual Play: A "Play Next Move" button allows you to manually play the next move after you analyze the board.
• Graphical User Interface (GUI): Provides a user-friendly interface instead of terminal-based interaction.
• New Feature – ESC Key: Allows the user to go back and select the playing color again.
• New Feature – Castling Support: Users can now indicate castling rights via a tick mark before each move.
• New Feature – Depth Slider: Users can now choose the depth for Stockfish's analysis via a slider. Note: More depth means longer processing time. The default depth is set to 15, which is recommended for a good balance between speed and accuracy.
• New Feature – Retry Mechanism: If the piece doesn’t move as expected, the tool will automatically retry the move up to 3 times —to ensure the move is successfully executed.
• Optimized Performance: The removal of external APIs results in lower latency and improved move execution speed.

---

Download

Get the latest release from GitHub Releases.

---

Prerequisites

Ensure the following software is installed:

• Python 3.10+
• Python Libraries:
  • mss (pip install mss)
  • Pillow (pip install Pillow)
  • pyautogui (pip install pyautogui)
  • onnxruntime (pip install onnxruntime)
  • tkinter (usually pre-installed with Python)
• Stockfish Chess Engine: Download here

---

Installation

1. Clone the Repository:

   git clone https://github.com/OTAKUWeBer/ChessPilot.git
   cd ChessPilot

2. Install Python Packages:

   pip install -r requirements.txt

3. Download the Chess Detection Model

   • Download the chessboard detection model from Zai-Kun’s Chess Pieces Detection
   • Rename it as chess_detection.onnx
   • Place this file inside the ChessPilot directory (same as main.py).

Linux Users

Install Stockfish via your package manager:

• Debian/Ubuntu:

  sudo apt install stockfish

• Arch Linux:

  paru/yay -S stockfish

• Fedora:

  sudo dnf install stockfish

Wayland Users (Hyprland, Sway)

If you're using a Wayland compositor (such as Hyprland or Sway), you'll need to install additional dependencies for screenshot functionality and screen resolution detection:

sudo pacman -S grim wayland-utils

Windows Users

1. Stockfish:
   • Download stockfish.exe from Stockfish, rename it to stockfish.exe, and place it inside the ChessPilot directory (same as main.py).

---

Usage

1. Run the Script:

   python main.py

2. Choose Your Color via the GUI:

   • Click White if playing as White.
   • Click Black if playing as Black.
   • Press ESC to go back and re-select your color.

3. Indicate Castling Rights:

   • Before each move, tick the checkbox for Kingside Castle or Queenside Castle, if applicable.

4. Choose Depth for Move:

   • Use the slider to select the desired depth for Stockfish's analysis. Note: The higher the depth, the more time it will take for Stockfish to compute the best move. The default depth is set to 15, which is suggested for optimal performance.

5. Play the Game:

   • Manual Mode: Click the "Play Next Move" button to manually analyze and execute your next move.
   • Auto Mode: Enable "Auto Play Moves" to let the script automatically play the next move after the opponent’s move without needing to click anything.

6. Retry Mechanism:

   • If the piece does not move as expected after executing a move (or castling move), the tool will automatically retry up to 3 times.
   • During each retry, the status will display messages like "Piece didn't move, retrying (1/3)".
   • Once the piece moves successfully, the tool will confirm with a success message, e.g., "Best Move: e2e4 executed successfully after retry (2/3)."

7. Tips for Best Performance:

   • Use 100% zoom for better accuracy in chessboard detection.
   • After selecting "Play next move" or enabling "Auto Play Moves," the script will:
     • Extract the FEN from the image.
     • Flip the board if playing as Black.
     • Retrieve the best move from Stockfish.
     • Automatically execute the move.
     • The best move will be displayed in the GUI.
     • For improved visibility, reposition the Tkinter window so the chessboard remains clear.
   • In Auto Mode, the script will automatically make your next move after the opponent’s move.

---

Disclaimer

🛑 Use at Your Own Risk: Using this tool in online chess games may lead to account bans.

---

License

This project is licensed under the MIT License.

---

Contributing

Contributions are welcome! Feel free to open issues or submit pull requests if you have suggestions or improvements.

---

Acknowledgments

• Thanks to Zai-Kun for creating the chessboard detection model and adding Wayland support.