diff --git a/rope/Coordinator.py b/rope/Coordinator.py
index c7ffda90..c35457ea 100644
--- a/rope/Coordinator.py
+++ b/rope/Coordinator.py
@@ -111,6 +111,9 @@ def coordinator():
             gui.set_video_slider_length(action[0][1])
             action.pop(0)
 
+        elif action[0][0] == "update_markers_canvas":
+            gui.update_markers_canvas()
+            action.pop(0)        
           
             
         else:
diff --git a/rope/GUI.py b/rope/GUI.py
index f90755e0..4b9e92e7 100644
--- a/rope/GUI.py
+++ b/rope/GUI.py
@@ -30,6 +30,7 @@ def __init__(self, models):
         self.title('Rope-Opal-03a')
         self.target_media = []
         self.target_video_file = []
+        self.media_file_name = []
         self.action_q = []
         self.video_image = []
         self.video_loaded = False
@@ -62,7 +63,6 @@ def __init__(self, models):
         self.widget = {}
         self.static_widget = {}
         self.layer = {}
-        
 
 
         self.json_dict =    {
@@ -1308,6 +1308,7 @@ def load_target(self, button, media_file, media_type):
         self.layer['markers_canvas'].delete('all')
         self.markers = []
         self.stop_marker = []
+        self.load_markers_json()
         self.add_action("markers", self.markers)
 
 
@@ -1488,6 +1489,7 @@ def temp_toggle_swapper(self, state):
             
     def toggle_rec_video(self):
         # Play button must be off to enable record button
+        self.save_markers_json()
         if not self.widget['TLPlayButton'].get():
             self.widget['TLRecButton'].toggle_button()
                 
@@ -1684,17 +1686,33 @@ def sort(e):
         
         # resize canvas
         else :
+            self.add_action("update_markers_canvas", self.markers)
 
-            self.layer['markers_canvas'].delete('all')
-            width = self.layer['markers_canvas'].winfo_width()-20-40-20
-            
-            for marker in self.markers:
-                position = 20+int(width*marker['frame']/self.video_slider.get_length())
-                marker['icon_ref'] = self.layer['markers_canvas'].create_line(position,0, position, 15, fill='light goldenrod')
-
+    def save_markers_json(self):
+        if len(self.markers) == 0 or len(self.media_file_name) == 0:
+            return
+        json_file_path = os.path.join(self.json_dict["source videos"], self.media_file_name[0] + "_markers.json")
+        # Save the markers to the JSON file
+        with open(json_file_path, 'w') as json_file:
+            json.dump(self.markers, json_file)
 
+    def load_markers_json(self):
+        if len(self.media_file_name) == 0:
+            return
+        json_file_path = os.path.join(self.json_dict["source videos"], self.media_file_name[0] + "_markers.json")
+        if os.path.exists(json_file_path):
+            # Load the markers from the JSON file
+            with open(json_file_path, 'r') as json_file:
+                self.markers = json.load(json_file)
+            self.add_action("update_markers_canvas", self.markers)
+
+    def update_markers_canvas(self):
+        self.layer['markers_canvas'].delete('all')
+        width = self.layer['markers_canvas'].winfo_width()-20-40-20
+        for marker in self.markers:
+            position = 20+int(width*marker['frame']/self.video_slider.get_length())
+            marker['icon_ref'] = self.layer['markers_canvas'].create_line(position,0, position, 15, fill='light goldenrod')
 
-                
     def toggle_stop(self):
         if self.stop_marker == self.video_slider.self.timeline_position:
             self.stop_marker = []
diff --git a/rope/VideoManager.py b/rope/VideoManager.py
index 45727d39..16da1cf0 100644
--- a/rope/VideoManager.py
+++ b/rope/VideoManager.py
@@ -150,6 +150,7 @@ def load_target_video( self, file ):
             self.r_frame_q = []             
             self.found_faces = []
             self.add_action("set_slider_length",self.video_frame_total-1)
+            self.add_action("update_markers_canvas", self.markers)
 
         self.capture.set(cv2.CAP_PROP_POS_FRAMES, self.current_frame)        
         success, image = self.capture.read()