Metin2 AI Vision Bot ⚔️💎

An advanced automation tool for Metin2 using YOLOv8 for real-time object detection and OpenCV for visual processing. This bot is designed to detect and farm stones (Metins) autonomously.

Disclaimer: This project is for educational purposes and to demonstrate computer vision applications. Use at your own risk.

🚀 Features

• AI-Powered Detection: Uses YOLOv8 to identify stones in real-time.
• Smart Logic: Prioritizes the nearest stone and avoids redundant clicks.
• Auto Skill Casting: Automatically dismounts, casts predefined skills, and remounts at set intervals.
• Auto Respawn: Detects character death and clicks the "Restart" button via template matching.
• GUI Control Panel: User-friendly interface to manage bot states and timings.

🧠 Pre-trained Models

This repo includes two specialized YOLOv8 models:

• best.pt: Optimized for Red Forest (Kızıl Orman) stones.
• best2.pt: Optimized for the Custom Map (snow_stone).

Switch between them by modifying MODEL_PATH in config.py.

🛠️ Configuration & Important Notes

• Skill Setup: Place your buff/utility skills on slots 1 and 2. The bot is configured to use these keys.
• Attack Duration: In the GUI, set the Attack Duration (s) (e.g., 5.0 or 7.0). This defines the time the bot waits before looking for the next stone after a click.
• Buff Interval: By default, the bot casts skills every 25 minutes (1500 seconds). You can change this by editing SKILL_INTERVAL in config.py.
• Admin Privileges: You MUST run your terminal/IDE as Administrator for the mouse and keyboard emulation (SendInput) to work within the game.
• Window Mode: The game must not be in full screen. Run it in windowed mode for the bot to function properly.

📈 How to Train for Custom Maps (Roboflow)

To adapt the bot to different maps:

1. Data Collection: Take 30-50 screenshots of the target stones from various angles. Reference examples are in Stones_img/.
2. Annotation: Create a project on Roboflow (free), upload images, and draw bounding boxes.
3. Training: Train a YOLOv8 model using Roboflow's cloud training or locally.
4. Deployment: Save the .pt weights in the project folder and update config.py.

For a detailed walkthrough, check out the original project foundation: jamnicki/metin2_vision_bot.

📦 Installation

1. Clone the repository:

   git clone https://github.com/muhammetsimssek/Metin2-AI-Vision-Bot.git
   cd Metin2-AI-Vision-Bot

2. Install dependencies:

   pip install -r requirements.txt

3. Launch the Bot:

   python main.py

---

Developed by Muhammet