diff --git a/app.py b/app.py
new file mode 100644
index 000000000..a0e1a6f39
--- /dev/null
+++ b/app.py
@@ -0,0 +1,46 @@
+from beam import task_queue, Output, Image, env, Volume
+import os
+import sys
+import requests
+if env.is_remote():
+     import folder_paths
+     sys.path.append(os.getcwd())
+     import main
+
+
+# Function to setup environment and load models
+def upload_models():
+    # Define the model URLs and their respective directories
+    model_urls = [
+      #put models here
+    ]
+    
+    # Download models if not already downloaded
+    for model_name, filename, directory in model_urls:
+        if not os.path.exists(f"{directory}/{filename}"):
+            print(os.getcwd())
+            os.system(f"wget -c https://huggingface.co/{model_name}/resolve/main/{filename} -P {os.path.join(os.getcwd() , directory)}")
+
+
+
+# Import functions or execute code from workflow_api.py
+
+# Step 1: Define environment setup and model loading
+@endpoint(
+    name="ComfyUI",
+    cpu=4,
+    memory="32Gi",
+    gpu="A10G",
+    image=Image(
+        python_version="python3.10",
+        python_packages=requirements.txt,
+        #commands=["pip3 install opencv-python"],
+    ),
+    volumes=[Volume(name="PUTVOLUMEHERE", mount_path="./VOLUMEPATH)],
+    #keep_warm_seconds=0,
+    #on_start=upload_models,
+)
+def generate_video(**inputs):
+  while(true):
+  main()
+