diff --git a/imagestuff.py b/imagestuff.py
new file mode 100644
index 0000000..56f3eb0
--- /dev/null
+++ b/imagestuff.py
@@ -0,0 +1,103 @@
+import tkinter as tk
+from PIL import Image, ImageTk
+import subprocess
+import time
+import cv2
+import pyautogui
+from screeninfo import get_monitors
+from mss import mss  
+
+def get_primary_monitor():
+    for m in get_monitors():
+        if m.is_primary:
+            return m
+    return None
+
+class ScreenshotSelector:
+    def __init__(self, root):
+        self.root = root
+        self.start_x = None
+        self.start_y = None
+        self.rect = None
+        self.temp_screenshot_path = '/tmp/temp_screenshot.png'
+        self.button_image_path = None
+
+        # Delay and capture
+        time.sleep(5)
+        self.capture_primary_display()
+
+        # Load the screenshot for display
+        self.image = Image.open(self.temp_screenshot_path)
+        self.tkimage = ImageTk.PhotoImage(self.image)
+
+        self.canvas = tk.Canvas(root, width=self.image.width, height=self.image.height)
+        self.canvas.pack()
+        self.canvas.create_image(0, 0, anchor='nw', image=self.tkimage)
+
+        # Bind mouse events
+        self.canvas.bind("<ButtonPress-1>", self.on_press)
+        self.canvas.bind("<B1-Motion>", self.on_drag)
+        self.canvas.bind("<ButtonRelease-1>", self.on_release)
+
+        # Bind key events
+        self.root.bind("<d>", self.on_d_press)
+
+    def capture_primary_display(self):
+        # Command to capture the primary display
+        cmd = ["gnome-screenshot", "-f", self.temp_screenshot_path]
+
+        # Execute the command
+        subprocess.run(cmd, check=True)
+
+    def on_press(self, event):
+        self.start_x = self.root.winfo_pointerx() - self.root.winfo_rootx()
+        self.start_y = self.root.winfo_pointery() - self.root.winfo_rooty()
+        self.rect = self.canvas.create_rectangle(self.start_x, self.start_y, self.start_x, self.start_y, outline='red')
+
+    def on_drag(self, event):
+        curX = self.root.winfo_pointerx() - self.root.winfo_rootx()
+        curY = self.root.winfo_pointery() - self.root.winfo_rooty()
+        self.canvas.coords(self.rect, self.start_x, self.start_y, curX, curY)
+
+    def on_release(self, event):
+        end_x = self.root.winfo_pointerx() - self.root.winfo_rootx()
+        end_y = self.root.winfo_pointery() - self.root.winfo_rooty()
+        self.capture_region(self.start_x, self.start_y, end_x, end_y)
+        self.root.destroy()
+
+    def capture_region(self, x1, y1, x2, y2):
+        region_screenshot = self.image.crop((x1, y1, x2, y2))
+        self.button_image_path = '/tmp/button_image.png'
+        region_screenshot.save(self.button_image_path)
+        print(f"Button image saved to {self.button_image_path}")
+
+    def on_d_press(self, event):
+        if self.button_image_path:
+            self.find_button_and_click(self.button_image_path)
+
+    def find_button_and_click(self, button_image_path):
+        # Take a current screenshot of the primary display
+        self.capture_primary_display()
+
+        # Load the current screenshot and button image
+        current_screenshot = cv2.imread(self.temp_screenshot_path, 0)
+        button_image = cv2.imread(button_image_path, 0)
+        w, h = button_image.shape[::-1]
+
+        # Template matching
+        res = cv2.matchTemplate(current_screenshot, button_image, cv2.TM_CCOEFF_NORMED)
+        min_val, max_val, min_loc, max_loc = cv2.minMaxLoc(res)
+        print(f"Match percentage: {max_val * 100:.2f}%")
+
+        # Click if a good match is found
+        if max_val > 0.8:  # Adjust this threshold as needed
+            click_x, click_y = max_loc[0] + w//2, max_loc[1] + h//2
+            pyautogui.click(click_x, click_y)
+            print("Clicked on the button.")
+        else:
+            print("No suitable match found.")
+
+if __name__ == "__main__":
+    root = tk.Tk()
+    ScreenshotSelector(root)
+    root.mainloop()