Update app.py

app.py

@@ -71,7 +71,7 @@ async def chat_with_ai(message):
     
     try:
         response = chat_client.chat_completion(
-            messages=[{"role": "system", "content": "You are a helpful voice assistant. Provide concise and clear responses to user queries."}] + chat_history,
+            messages=[{"role": "system", "content": "You are a helpful voice assistant. Provide concise and clear responses to user queries.Always try to respond in English if the query is in any other language."}] + chat_history,
             max_tokens=800,
             temperature=0.7
         )
@@ -98,25 +98,62 @@ def transcribe_and_chat(audio):
     return response, audio_path
 
 def create_demo():
-    with gr.Blocks() as demo:
-        gr.Markdown("# AI Voice Assistant")
+    with gr.Blocks(theme=gr.themes.Soft()) as demo:
+        gr.Markdown(
+        """
+        # 🎙️ AI Voice Assistant
+        Welcome to the AI Voice Assistant! Speak your question or command, and I'll respond with both text and voice.
+        """
+        )
 
         with gr.Row():
             with gr.Column(scale=1):
-                audio_input = gr.Audio(type="filepath", label="Press 'Record' to Speak")
+                audio_input = gr.Audio(
+                    type="filepath",
+                    label="🎤 Press 'Record' to Speak",
+                    source="microphone"
+                )
+
+            with gr.Column(scale=2):
+                chat_output = gr.Chatbot(
+                    label="Conversation",
+                    height=300
+                )
+                audio_output = gr.Audio(
+                    label="🔊 AI Voice Response",
+                    autoplay=True
+                )
 
-            with gr.Column(scale=1):
-                chat_output = gr.Textbox(label="AI Response")
-                audio_output = gr.Audio(label="AI Voice Response", autoplay=True)
+        with gr.Row():
+            clear_button = gr.Button("🧹 Clear Conversation")
 
-        def process_audio(audio):
+        def process_audio(audio, history):
             logging.info(f"Received audio: {audio}")
             if audio is None:
-                return "No audio detected. Please try recording again.", None, None
-            response, audio_path = transcribe_and_chat(audio)
-            logging.info(f"Response: {response}, Audio path: {audio_path}")
-            return response, audio_path, None  # Return None to clear the audio input
+                return history + [["No audio detected. Please try recording again.", None]], None, None
+            user_message, audio_path = transcribe_and_chat(audio)
+            assistant_response = f"You said: {user_message}\n\nMy response: {audio_path[0]}" if isinstance(audio_path, tuple) else f"You said: {user_message}\n\nMy response: {audio_path}"
+            history.append((user_message, assistant_response))
+            logging.info(f"Response: {assistant_response}, Audio path: {audio_path}")
+            return history, audio_path, None  # Return None to clear the audio input
+
+        def clear_conversation():
+            global chat_history
+            chat_history = []
+            return None, None, None
+
+        audio_input.change(
+            process_audio,
+            inputs=[audio_input, chat_output],
+            outputs=[chat_output, audio_output, audio_input]
+        )
 
+        clear_button.click(
+            clear_conversation,
+            outputs=[chat_output, audio_output, audio_input]
+        )
+
+        # Custom JavaScript for auto-submission and audio playback
         demo.load(None, js="""
             function() {
                 document.querySelector("audio").addEventListener("stop", function() {
@@ -145,8 +182,6 @@ def create_demo():
             }
         """)
 
-        audio_input.change(process_audio, inputs=[audio_input], outputs=[chat_output, audio_output, audio_input])
-
     return demo
 
 # Launch the Gradio app