feat: Enable MCP (#4)

- feat: Enable MCP (3b9399cb5a39200430a8a9e0d3aa5139ead2eb42)


Co-authored-by: Apolinário from multimodal AI art <multimodalart@users.noreply.huggingface.co>

app.py

@@ -230,6 +230,19 @@ def get_video_duration(video_path):
 @torch.inference_mode()
 @torch.no_grad()
 def synthesize_video_with_audio(video_file, caption, cot):
+    """
+    Synthesizes audio for a given video using text captions and chain-of-thought descriptions.
+    
+    Args:
+        video_file: Path to the input video file
+        caption: Optional text caption describing the desired audio
+        cot: Optional chain-of-thought description for enhanced audio generation
+        
+    Returns:
+        Generator yielding tuples of (status_message, output_video_path) where output_video_path
+        is None during processing and contains the path to the final video with synthesized audio
+        when complete
+    """
     yield "⏳ Extracting Features…", None
     video_path = video_file
     if caption is None:
@@ -335,9 +348,6 @@ demo = gr.Interface(
 )
 
 if __name__ == "__main__":
-    demo.queue().launch(share=True)
-
-demo.launch(share=True)
-
-
+    demo.queue().launch(share=True, mcp_server=True)
 
+demo.launch(share=True, mcp_server=True)