Update app.py

app.py

@@ -243,12 +243,15 @@ with gr.Blocks(title="Orpheus Text-to-Speech") as demo:
         outputs=[text_input, audio_output]
     )
 
-# Create FastAPI app and mount Gradio
-demo = demo.queue()
+# Enable queuing for Gradio
+queue_demo = demo.queue()
+
+# Create FastAPI app and mount Gradio ASGI app
 app = FastAPI()
-app.mount("/", demo)
+app.mount("/", queue_demo.app)  # mount the ASGI app, not the Blocks object itself
 
-# WebSocket TTS endpoint\@app.websocket("/ws/tts")
+# WebSocket TTS endpoint
+@app.websocket("/ws/tts")
 async def websocket_tts(websocket: WebSocket):
     await websocket.accept()
     try:
@@ -257,10 +260,8 @@ async def websocket_tts(websocket: WebSocket):
             data = json.loads(msg)
             text = data.get("text", "")
             voice = data.get("voice", VOICES[0])
-            # Generate audio for the chunk
             _, audio = generate_speech(text, voice, 0.7, 0.95, 1.1, 1200)
-            # Stream audio in 0.1s chunks
-            chunk_size = 2400  # 24000 Hz -> 2400 samples = 0.1s
+            chunk_size = 2400  # 0.1s at 24kHz
             for i in range(0, len(audio), chunk_size):
                 chunk = audio[i:i+chunk_size]
                 await websocket.send_bytes(chunk.tobytes())
@@ -268,7 +269,7 @@ async def websocket_tts(websocket: WebSocket):
     except WebSocketDisconnect:
         print("Client disconnected from /ws/tts")
 
-# Launch if run directly
+# Launch when run directly
 def main():
     import uvicorn
     uvicorn.run("app:app", host="0.0.0.0", port=7860)