Update ui/voice_agent_ui.py

ui/voice_agent_ui.py

@@ -139,7 +139,7 @@ def create_voice_agent_ui(agent):
 
     # MAIN COMMAND PROCESSOR
     async def process_audio_command(audio_file, text_command, files_list):
-        """Process voice + text commands - FIXED for standard Gradbot format"""
+        """Process voice + text commands - FIXED for tuple format"""
 
         # Step 1 — Identify user command
         if audio_file and not text_command:
@@ -165,13 +165,12 @@ def create_voice_agent_ui(agent):
             # Call agent (non-streaming)
             final_answer, thoughts = await agent.execute(cmd, files_list)
 
-            # Convert AgentThought objects to standard Gradio Chatbot format
-            # Format: List of tuples [(user_msg, bot_msg), ...]
+            # Convert AgentThought objects to Gradio Chatbot tuple format
+            # MUST be: [(user_msg, bot_msg), (user_msg, bot_msg), ...]
+            # Each tuple is ONE chat exchange
             messages = []
-            current_user_msg = "🎯 Task Progress:"
-            current_bot_msgs = []
             
-            for t in thoughts:
+            for i, t in enumerate(thoughts):
                 # Handle both AgentThought objects and dicts
                 if hasattr(t, "type"):
                     t_type = t.type
@@ -202,12 +201,10 @@ def create_voice_agent_ui(agent):
                     icon = "✅"
                     title = "Answer"
 
-                # Add to current bot messages
-                current_bot_msgs.append(f"{icon} **{title}**\n\n{t_content}")
-
-            # Create single chat entry with all thoughts
-            if current_bot_msgs:
-                messages.append((current_user_msg, "\n\n---\n\n".join(current_bot_msgs)))
+                # Create a tuple for each thought: (user_msg, bot_msg)
+                user_msg = f"Step {i+1}"
+                bot_msg = f"{icon} **{title}**\n\n{t_content}"
+                messages.append((user_msg, bot_msg))
 
             # Show results
             yield messages, "📊 Generating voice response...", final_answer, None, None
@@ -231,8 +228,8 @@ def create_voice_agent_ui(agent):
             import traceback
             err_msg = f"⚠️ Error: {str(e)}\n\nTraceback:\n{traceback.format_exc()}"
             
-            # Error message in standard format
-            error_messages = [("❌ Error occurred", f"**Error**\n\n{str(e)}")]
+            # Error message as tuple
+            error_messages = [("Error", f"❌ **Error**\n\n{str(e)}")]
             yield error_messages, f"❌ Error: {str(e)}", err_msg, None, None
 
     # CONNECT EVENTS