diff --git a/ml/ml_play_manual.py b/ml/ml_play_manual.py
new file mode 100644
index 0000000..5c612fe
--- /dev/null
+++ b/ml/ml_play_manual.py
@@ -0,0 +1,38 @@
+import random
+import pygame
+
+
+class MLPlay:
+    def __init__(self):
+        print("Initial ml script")
+
+    def update(self, scene_info: dict, keyboard:list=[], *args, **kwargs):
+        """
+        Generate the command according to the received scene information
+        """
+        # print("AI received data from game :", json.dumps(scene_info))
+        # print(scene_info)
+        actions = []
+
+        if pygame.K_w in keyboard or pygame.K_UP in keyboard:
+            actions.append("UP")
+        elif pygame.K_s in keyboard or pygame.K_DOWN in keyboard:
+            actions.append("DOWN")
+
+        elif pygame.K_a in keyboard or pygame.K_LEFT in keyboard:
+            actions.append("LEFT")
+        elif pygame.K_d in keyboard or pygame.K_RIGHT in keyboard:
+            actions.append("RIGHT")
+        else:
+            actions.append("NONE")
+
+        return actions
+
+
+
+    def reset(self):
+        """
+        Reset the status
+        """
+        print("reset ml script")
+        pass
diff --git a/ml/ml_play_template.py b/ml/ml_play_template.py
new file mode 100644
index 0000000..02efb26
--- /dev/null
+++ b/ml/ml_play_template.py
@@ -0,0 +1,21 @@
+import random
+class MLPlay:
+    def __init__(self):
+        print("Initial ml script")
+
+    def update(self, scene_info: dict, *args, **kwargs):
+        """
+        Generate the command according to the received scene information
+        """
+        # print("AI received data from game :", json.dumps(scene_info))
+        # print(scene_info)
+        actions = ["UP", "DOWN", "LEFT", "RIGHT"]
+
+        return random.sample(actions, 1)
+
+    def reset(self):
+        """
+        Reset the status
+        """
+        print("reset ml script")
+        pass