Make an autoclicker

Budget: 12000 UAH Deadline: 7 days

Good day, I have experience in developing autoclickers and similar projects. I have a few questions regarding the provided specifications, please write to me privately for clarification.

Looking forward to future collaboration!

Budget: 9500 UAH Deadline: 7 days

Hello, I can take on the job. I have a lot of experience working with OpenCV, TensorFlow/Keras, PyTorch, dlib, Tesseract OCR, MediaPipe, PyAutoGUI, NumPy. I used to write scripts based on computer vision for various tasks, recently about 15, and I worked with recognizing license plates, people, text, bypassing captchas, etc. I'm curious why you need ChatGPT for decision-making; I don't quite understand yet, but I will integrate it if needed. Also, I'm interested in how global the task is because if it's not too complicated, I can easily do it for 2000 UAH - 4000 UAH in 1 day.

(Recording and playback of user actions) - this is very easy, especially if it involves video or screenshots, then about 20 minutes of work, give or take, if the scale is not large.

(I will set the frequency and duration of clicks) in 2 minutes using win32api, win32con, or pyautogui.

(Analyzing the screen state before performing actions) - this is generally a piece of cake. We focus on the screen we are working with through hwnd:
app = Application().connect(handle=hwnd)
window = app.window(handle=hwnd)
window.set_focus()

Then we work with the cv2 module for image recognition.

(Reacting to interface changes (for example, the appearance of an error window or the completion of a download)) - we also process this through the cv2 module. I skipped ads, so I reacted to errors. If there are 5 such pop-up windows, I can handle it in 10 minutes easily.

(Using hotkeys to turn on/off) is very easy to implement; we simply handle the key press with an IF code or launch a second thread and process it depending on the task.

(Logging actions for subsequent analysis) - I can output to the console easily in any color, and I can also save logs if needed; nothing complicated.

(Interacting with the neural network for decision-making based on the current situation) - this is where there is some complexity because I don't fully understand the specifics and scope, so I need more details. However, for me, any complexity is not a problem; we need to apply the already trained neural network. If needed, we can train it. But training is a bit longer process; it depends on the task.

Write to me, consult me on everything, and I will do it all.

I also can't set a lower budget because Freelancehunt doesn't allow it; I registered today and don't understand why yet. Expand