Thanks : )

@@CheesyAI good job!

most of csgo cheats written on C family languages, like C++, C#, because they are fast

use yolov8.2, it has grab screen realtime

cheater

No neither of those anticheats detect mouse_event. For smoothing you would need to create a lerp fucntion for each mouse_event. How ever something like this is far to slow to be used as a cheat. If youre goint to do this i would switch to the mss library over pyautogui.

Yoo were u able to make it?

Try a fork of this project in the video: github.com/JiaPai12138/Crossfirer , download AI_M_BOT vx.x.x and install. Tell me if you want English version. Basically select 1 and drag your csgo window (please make it windowed) a bit, then wait for the small window on top left of your screen. Press 1 or 2 to let program control your mouse... +) fastest screenshot method on windows (thank to another guy) +) multiprocessing +) region is 2/3 of height * 2/3 of height +) could be 60+ fps on gtx1660-ti -) react slow on 35 minus fps machine -) mouse move unlike human

@@killmatt5574 Crossfirer may have a bug, if you run it as admin it crashes, if you dont, it runs well but auto mouse movement doesnt work

@@xGabz Crashed? Which error message is shown?

@@killmatt5574 It does not give an error, for a fraction of a second it shows the menu and then, it will just close the terminal (btw im using yolov3 tiny, i don't think it matters)

@@xGabz Will update soon. Do you have ndivia card and driver installed?

I don't think that there would be much changing. I assume that there would be a button on the controller that you might wanna hold to activate the lock-on mechanism. In the code you'd only change this button and the rest of it would work.

Using windows api, specifically mouse_event

@@MadBuddhaAbuser I just want my mouse to rotate to the right continuously during the game. what script do i have to use in python

Compile it

It is yolov4, you can download here: github.com/AlexeyAB/darknet/wiki/YOLOv4-model-zoo

i think its Atom ide but i am not sure

how are you unable to find cmake, the number 1 option for building c++ projects, also why do you need cmake?

Write your own pirate drivers

Yes but detected

Good point thanks

@@CheesyAI are you developing this further? I'm currently collecting a dataset of players on multiple maps, 300 x 300 centre of screen as a low fov silent aimbot

@@lightningblizerd have you created anything would love to try it

its pretty simple

remove lines 78-81

Alright, let's cut to the chase. Your take on Python being slow is about as accurate as a broken clock. You're right twice a day, but let's face it, that's not a winning average. You're screaming from the rooftops that Python is a tortoise, but you're ignoring the C libraries - the rocket boosters, if you will - strapped to its back. It's like saying a train is slow while ignoring the fact it's running on jet fuel. Get your facts straight, buddy. Honestly, your understanding of game hacking is so off base, it's in a different ballpark. You're throwing around advice like you're some kind of coding guru, but the truth is, you couldn't find your way out of a paper bag with a map and a flashlight. And your tirade against Python for using libraries? That's like blaming a car for having wheels. They're not a crutch; they're the whole point. They're the tools that make the magic happen. But with your logic, we should all be reinventing the wheel every time we code. So, before you start spouting off about things you clearly know zip about, do us all a favor and zip it. Because right now, the only thing you're hacking is the truth. But by all means, keep yapping. It's entertaining to watch you make a fool of yourself.

@@user-pf9jv1fl2n I got 3 word for you , I use rust. And btw u can run your on test’s have fun and keep on learning .

It's a screenshot that is fed to an neural network to get the coordinates of players. It's much slower and less effective than memory reading but also harder to detect.

Each frame gets fed to the NN and analyzed by the model. The model defines key areas like center of the screen, x-hair pos, and through time it learns if certain pixels are objects. You need to train the model with many frames to be confident. Of course this is going to be slower than injecting into a running executable.

Because he is using pre-trained model using coco dataset

@@killmatt5574 no, cpcp dataset cannot have data about this game. I think he does not want to share it

@@Nguyenthao-oj9ku I am training for another game. It is hell of a work...

@@killmatt5574 Can you please tell me how can you let the game window as the input for Yolov?

@@Nguyenthao-oj9ku Do screenshot. I don't see a way that the game window can be directly used as input

No

[ln[i[0] - 1] ->[ln[i - 1]

@@Moon-D0G Thank you brother

Yes, should work on even potato

these kind of “hacks” only works if fps > 100 and with customly trained model. I got 200+ on a 3060 . Of course i dont use this, but this was a good beginner’s initial challenge.( small tip for you : reduce the frame size, and change backend of the model to fp16 since you said you have rtx2060, gtx wont work with fp16 though). btw I use C++.

I think that might not be much faster since yolo is written in C and the bottleneck here is that detection.

yeah another way is to compile with nuitka...if works

@@cosmilitebar6772 yolo is not the fastest. This is a god project what can be much faster. And the problem is in screen capture , in optimization and in yolo

Undetected for almost 20 years

to get past kernel anti cheat just stream the game from another system

Run the script

You may have to pip install numpy in your terminal, than import numpy

For esp u need to inject and can't just use ai bots

@@NemralXamsimus You dont need to inject, you could just create a top most window and use the ai's findings to draw boxes, but this would only show you what you already can see. For real esp you would need to use pymem and read the memory positions of all enemies.

@@MadBuddhaAbuser well yes 🥸

@@MadBuddhaAbuser how?

Yes, with human shapes as players

i dont kno, try

Or maybe you just learn how to play

change line 20 a bit to the emulator you are using

Bro i have ff aimbot

You can training your own dataset

@@killmatt5574 cool