Version: 2.1.0 β
Last Updated: April 27, 2025
Update main.py and model weights in order for script to be compatible with TGE update!
BlumClicker is an automation tool for Blum's Drop Game on Telegram, powered by YOLOv11 computer vision technology for precise snowflake detection and interaction. The tool features a graphical interface, extensive customization options, and intelligent gameplay automation.
Русская версия здесь.
- Features
- System Requirements
- Installation
- Usage
- Configuration
- Hotkeys
- Troubleshooting
- Advanced Features
- FAQ
- Contributing
- Contact
- Window Recognition - Automatically detects and focuses on the Telegram game window
- YOLOv11 Object Detection - State-of-the-art visual recognition for snowflake targets
- OCR Technology - Detects and clicks the "Play" button for automatic game restart
- Real-time Statistics - Monitor performance metrics, click counts, and system resource usage
- Rich GUI Interface - Visualize bot activity with detailed panels and real-time updates
- Configurable Parameters - Customize delays, FPS limits, detection thresholds and more
- Debug Visualization - Optional visual feedback showing detection boundaries and confidence scores
- Hotkey Controls - Convenient keyboard shortcuts for all major functions
- Operating System: Windows 10 (64-bit)
- CPU: Intel Core i5 or AMD Ryzen 5 equivalent
- RAM: 8GB
- GPU: NVIDIA GPU with CUDA support (tested with CUDA 11.8 and 12.5)
- Python: Version 3.12.x
- Storage: 2GB free space
- Software: Telegram Desktop application
- Operating System: Windows 10/11 (64-bit)
- CPU: Intel Core i7/i9+ or AMD Ryzen 7/9+
- RAM: 16GB+
- GPU: NVIDIA GPU with CUDA support (tested with CUDA 11.8 and 12.5)
- Python: Version 3.12.x
- Storage: 5GB+ storage
-
Clone the Repository
git clone https://github.com/fedyacpp/BlumClicker.git cd BlumClicker -
Install Required Python Packages
For GPU users (recommended for best performance):
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121 pip install -r requirements.txt
For CPU users:
pip install torch torchvision torchaudio pip install -r requirements.txt
-
Download Model File
Download the pre-trained YOLO model file:
- best.pt
- Place it in the BlumClicker root directory
-
Verify CUDA Availability (GPU users only)
python isCudaAvailable.py
-
Launch BlumClicker
python main.py
-
Start the Application
python main.py
-
Window Detection
- The bot will automatically detect your Telegram window
-
Controls
- Use the hotkeys to control the bot (see Hotkeys section)
- The main interface displays real-time statistics and status information
- Access settings via CTRL+W to customize behavior
-
Operation Modes
- Auto-Play Mode: Automatically restarts the game when it ends
- Manual Mode: Requires manual restart of games
- Debug Mode: Shows visual detection feedback
BlumClicker offers extensive configuration options through its settings panel (CTRL+W):
- Delay Between Clicks - Time to wait between consecutive clicks (seconds)
- Delay Before Click - Time to wait before performing a click (seconds)
- FPS Limit - Maximum frames processed per second
- Retry Count - Number of retry attempts for operations
- Auto-Play - Automatically start new games
- Debug Window - Show visual detection feedback
- Click All Bombs - Click on all detected objects, not just targets
- CPU Mode - Force CPU-only processing (disable GPU)
- Sound Effects - Enable/disable sound feedback for clicks
- Model Path - Location of the YOLO model file
- Model Reload - Ability to hot-swap models during runtime
| Hotkey | Function |
|---|---|
| CTRL+Q | Exit the application |
| CTRL+X | Pause/Resume bot operation |
| CTRL+W | Open settings panel |
| CTRL+D | Toggle debug visualization |
| CTRL+R | Reload the model |
| CTRL+F | Re-detect Telegram window |
| CTRL+A | Toggle auto-play mode |
| CTRL+S | Toggle sound effects |
| Problem | Solution |
|---|---|
| Window not detected | Ensure Telegram is running and visible; try CTRL+F to re-detect |
| CUDA errors | Verify GPU and CUDA installation; switch to CPU mode if needed |
| OCR not working | Ensure EasyOCR dependencies are installed properly |
| Inaccurate detection | Ensure the model file is correctly loaded and running on supported GPU |
| Permission errors | Run as administrator for proper window access |
- Check
bot_log.txtfor detailed error information - Enable debug mode (CTRL+D) to visualize detection performance
- Use the settings panel to view system resource usage and last errors
You can use your own trained YOLO models by selecting them in the settings panel:
- Train a custom YOLOv11 model
- Place the .pt file in an accessible location
- Set the model path in settings and reload
- GPU Acceleration: Enabled by default when available
- Memory Management: Configure with appropriate FPS limits
- CPU Mode: Available for systems without compatible GPUs
Q: Is this against Telegram's terms of service?
A: Automation tools may violate Telegram's terms of service. Use at your own discretion.
Q: Can I use multiple instances for different games?
A: Technically yes, but it won't work as supposed right out of the box.
Q: Does BlumClicker work on MacOS/Linux?
A: Currently, it's Windows-only due to the specific window handling mechanisms.
Q: How accurate is the detection?
A: With YOLOv11 and proper setup, detection accuracy is typically over 95%.
Q: Will my account get banned?
A: Use at your own risk. The tool tries to simulate human-like behavior, but no guarantees.
Contributions are welcome! Please follow these steps:
- Fork the repository
- Create a feature branch
- Make your changes
- Submit a pull request
- GitHub Issues: Report bugs or request features
- Email: [email protected]
- Telegram: @fedyacpp
Disclaimer: This tool is for educational purposes only. I'm not responsible for any misuse or violations of terms of service that may result from using this software.