Overview

Sunone Aimbot is an AI-powered aim bot for first-person shooter games. It leverages the YOLOv8 and YOLOv10 models, PyTorch, and various other tools to automatically target and aim at enemies within the game. The AI model in repository has been trained on more than 30,000 images from popular first-person shooter games like Warface, Destiny 2, Battlefield (all series), Fortnite, The Finals, CS2 and more.

[!WARNING] Use it at your own risk, we do not guarantee that you may be blocked!

[!NOTE] This application only works on Nvidia graphics cards. AMD support is testing. See AI_enable_AMD option. The recommended graphics card for starting and more productive and stable operation starts with the rtx 20 series.

Requirements

Before you get started, make sure you have the following prerequisites installed and pay attention to the versions in Tested Environment block, this may cause errors in launching the aimbot.

• Information in English:
  • Auto-Install guide
  • Self install guide
  • Questions and answers
  • Arduino setup
  • Discord server
    
    
• Информация на Русском языке:
  • Инструкция по автоматической установке
  • Инструкция по установке в ручную
  • Вопросы и ответы
  • Установка ардуино
  • Discord сервер
    
    
• To launch the aimbot after all installations, start run_ai.bat or type py run.py.

Tested Environment

The Sunone Aimbot has been tested on the following environment:

Options

The behavior of the aimbot can be configured via the config.ini file. Here are the available options:

Object Search window resolution:

• detection_window_width int: Horizontal resolution of the object search window.
• detection_window_height int: Vertical resolution of the object search window.

Bettercam capture method:

• Bettercam_capture bool: Use Bettercam to capture images from the screen.
• bettercam_capture_fps int: Specific fps value for screen capture.
• bettercam_monitor_id int: Id of the monitor from which the images will be captured.
• bettercam_gpu_id int: Id of the GPU to be used for image capture

Obs capture method:

• Obs_capture bool: Use Obs to capture images from the screen.
• Obs_camera_id str or int: auto or number of Virtual Camera ID.
• Obs_capture_fps int: Specific fps value for screen capture.

Aim:

• body_y_offset float: Allows correction of y coordinates inside the body detected box if head is not detected.
• hideout_targets bool: Allows shooting at targets on the range (for example in warface on the polygon or in aimlabs).
• disable_headshot bool: Disable head targerting.
• disable_prediction bool: Disable target position prediction.
• prediction_interval float: The time frame used to estimate an object's future position based on its current motion.
• third_person bool: Turn on the third-person game mode. (sunxds_0.5.7+)

Hot keys:

• The names of all the keys are here. Type None is empty button.
• hotkey_targeting str: Aiming at the target. Supports multi keys, for example hotkey_targeting = RightMouseButton,X2MouseButton
• hotkey_exit str: Exit.
• hotkey_pause str: Pause AIM.
• hotkey_reload_config str: Reload config.

Mouse:

• mouse_dpi int: Mouse DPI.
• mouse_sensitivity float: Aim sensitivity.
• mouse_fov_width int: The current horizontal value of the viewing angle in the game.
• mouse_fov_height int: The current vertical value of the viewing angle in the game.
• mouse_min_speed_multiplier float: Minimum mouse movement speed multiplier.
• mouse_max_speed_multiplier float: Maximum mouse movement speed multiplier.
• mouse_lock_target bool: True: Press once to permanently aim at the target, press again to turn off the aiming. False: Hold down the button to constantly aim at the target.
• mouse_auto_aim bool: Automatic targeting.
• mouse_ghub bool: Uses Logitech GHUB exploit for mouse input. Install this version.
• mouse_rzr bool: For Razer mouses. Install synapse 3.

Shooting:

• auto_shoot bool: Automatic shooting. (For some games need arduino).
• triggerbot bool: Automatic shooting at a target if it is in the scope, requires the mouse_auto_shoot option enabled, and aiming will also be automatically turned off.
• force_click bool: Shooting will be performed even if the sight is not located within the object.
• bScope_multiplier float: The multiplier of the target trigger size.

Arduino:

• arduino_move bool: Sends a command to the arduino to move the mouse.
• arduino_shoot bool: Sends a command to the arduino to fire with the mouse.
• arduino_port str: Arduino COM port. Use COM1 or COM2 ... or auto.
• arduino_baudrate int: Custom Arduino baudrate.
• arduino_16_bit_mouse bool: Send 16 bit data to the arduino port to move the mouse.

AI:

• AI_model_name str: AI model name.
• AI_model_image_size int: AI model image size.
• AI_conf float: How many percent is AI sure that this is the right goal.
• AI_device int or str: Device to run on, 0, 1... or cpu.
• AI_enable_AMD bool: Enable support Amd GPUs. Install ROCm, Zluda and PyTorch. See AMD docs.
• AI_mouse_net bool: Use a neural network to calculate mouse movements. See this repository.

Overlay:

• show_overlay bool: Enables the overlay. It is not recommended for gameplay, only for debugging.
• overlay_show_borders bool: Displaying the borders of the overlay.
• overlay_show_boxes bool: Display of boxes.
• overlay_show_target_line bool: Displaying the line to the target.
• overlay_show_target_prediction_line bool: Displaying the predictive line to the target.
• overlay_show_labels bool: Displaying label names.
• overlay_show_conf bool: Displaying the label names as well as the confidence level.

Debug window:

• show_window bool: Shows the OpenCV2 window for visual feedback.
• show_detection_speed bool: Displays speed information inside the debug window.
• show_window_fps bool: Displays FPS in the corner.
• show_boxes bool: Displays detectable objects.
• show_labels bool: Displays the name of the detected object.
• show_conf bool: Displays object confidence threshold for detection.
• show_target_line bool: Shows the mouse finishing line.
• show_target_prediction_line bool: Show mouse prediction line.
• show_bScope_box bool: Show the trigger box for auto shooting.
• show_history_points bool: Show history points.
• debug_window_always_on_top bool: The debug window will always be on top of other windows.
• spawn_window_pos_x int: When the debugging window starts, it takes the x position.
• spawn_window_pos_y int: When the debugging window starts, it takes the y position.
• debug_window_scale_percent int: Adjusts the size of the debug window.
• The names of the debugging window can be written in the file window_names.txt they will be randomly selected.

AI Models

• *.pt: Default AI model.
• *.onnx: The model is optimized to run on processors.
• *.engine: Final exported model, which is faster than the previous two.

Export .pt model to .engine

1. All commands are executed in the console window:
2. First, go to the aimbot directory using the command:

cd C:\Users\your_username\downloads\sunone_aimbot-main

3. Then export the model from the .pt format in .engine format.

yolo export model="models/sunxds_0.5.6.pt" format=engine device=0 imgsz=640 half=True

• model="model_path/model_name.pt": Path to model.
• format=engine: TensorRT model format.
• half=true: Use Half-precision floating-point format.
• device=0: GPU id.
• workspace=8: GPU max video memory.
• verbose=False: Debug stuff. Convenient function, can show errors when exporting.

4. Each model has its own image size with which it was trained, export only with the image size with which it was published.

Notes / Recommendations

• Limit the maximum value of frames per second in the game in which you will use it. And also do not set the screen resolution to high. Do not overload the graphics card.
• Do not set high graphics settings in games.
• Limit the browser (try not to watch YouTube while playing and working AI at the same time, for example (of course if you don't have a super duper graphics card)) and so on, which loads the video card.
• Try to use TensorRT for acceleration. .pt model is good, but does not have as much speed as .engine.
• Turn off the cv2 debug window, this saves system resources.
• Do not increase the object search window resolution, this may affect your search speed.
• If you have started the application and nothing happens, it may be working, close it with the F2 key and change the show_window option to True in the file config.ini to make sure that the application is working.

Support the project

I will post new models here.

License

This project is licensed under the MIT License. See LICENSE for