Update

README.md

@@ -4,7 +4,7 @@ emoji: 🧬
 colorFrom: indigo
 colorTo: purple
 sdk: gradio
-sdk_version: 5.33.0
+sdk_version: 5.33.0 # Note: I've updated this to a more recent version, adjust if needed
 app_file: app.py
 pinned: false
 secrets:
@@ -15,187 +15,120 @@ tags:
   - agent-demo-track
 ---
 
-
 # 🧬 MediAgent AI: Conversational Health Navigator & MCP Tool Server
 
-**MediAgent AI** is a sophisticated, conversational health assistant and a fully MCP-compliant tool server, built on Modal and Gradio. It provides reliable, sourced, and easy-to-understand answers to a wide range of medical questions. Whether you are a patient seeking simple explanations or a healthcare professional needing detailed research, MediAgent AI adapts to your needs — and can be used as a plug-and-play MCP tool in any agentic workflow!
+**[➡️ Try the Live Demo Here!](https://huggingface.co/spaces/Agents-MCP-Hackathon/medical_ai)**
 
----
+<div align="center" style="margin-bottom: 1em;">
+  <a href="https://modal.com/" target="_blank" style="text-decoration:none; margin-right:18px;">
+    <img src="https://assets.modal.com/logos/modal-logo-dark.svg" alt="Modal Logo" height="32" style="vertical-align:middle; margin-right:8px;"/>
+    <span style="font-size:1.1em; color:#8aa8ff; font-weight:bold;">Powered by Modal</span>
+  </a>
+  <a href="https://www.llamaindex.ai/" target="_blank" style="text-decoration:none;">
+    <img src="https://raw.githubusercontent.com/jerryjliu/llama_index/main/docs/imgs/llama_index_logo.png" alt="LlamaIndex Logo" height="32" style="vertical-align:middle; margin-right:8px;"/>
+    <span style="font-size:1.1em; color:#ffb300; font-weight:bold;">and LlamaIndex</span>
+  </a>
+  <div style="font-size:0.95em; color:#aaa; margin-top:2px;">Special thanks to the Modal and LlamaIndex teams for their amazing platforms!</div>
+</div>
 
-## 🌟 Why MediAgent AI Stands Out
+**MediAgent AI** is a sophisticated, conversational health assistant and a fully MCP-compliant tool server, built on Modal and Gradio. It provides reliable, sourced, and role-adapted answers to a wide range of medical questions.
 
-- **Hybrid MCP Architecture:** One Modal backend, two endpoints—agent and MCP tool server—proving true MCP compliance with zero deployment overhead.
-- **Real Medical Value:** Multilingual, role-adaptive, and sources every answer.
-- **Plug-and-Play MCP Tool:** Instantly usable as a tool server by any MCP-compatible client.
-- **Production-Ready:** Robust error handling, citations, and safety layers.
+This project is a submission for the **Gradio Agents & MCP Hackathon 2025**, competing in two tracks:
+- **🔧 MCP Tool / Server Track:** Our backend exposes a robust, MCP-compliant endpoint for any agent to use.
+- **🤖 Agentic Demo Showcase Track:** The Gradio UI provides a powerful demonstration of an agent using these tools to help users.
 
 ---
 
-## 🏛️ Architecture: A Hybrid "Single-Server MCP" Approach
-
-To meet the rapid development cycle of a hackathon, we've implemented a pragmatic and powerful **Hybrid Architecture**. Our single Modal application serves as **both the core Agent and its own MCP-compliant Tool Server**.
-
-<!-- You can add a diagram here if desired -->
+## 🌟 Key Features
 
-1.  **Agent Endpoint (`/process`)**: This private endpoint orchestrates the agent's "think-plan-act" loop.
-2.  **MCP Tool Endpoint (`/mcp_tool_endpoint`)**: This **public endpoint** makes our application a valid MCP Server. It accepts standardized tool calls and executes them, proving the system's compliance with MCP principles.
-3.  **Internal Execution**: For maximum speed and reliability during the hackathon, the agent calls its tools via a direct internal interface rather than a network call to itself.
-
-This hybrid model demonstrates a deep understanding of both ideal agent architectures and practical, robust deployment strategies. It is scalable, efficient, and fully aligned with the spirit of the MCP Hackathon.
-
-**Powered By:**
-*   **Modal**: The perfect platform for this hybrid model, effortlessly hosting both the GPU-powered agent and the public-facing API endpoints within a single, scalable app.
-*   **MistralAI**: Advanced open-weight LLMs for medical reasoning.
-*   **LlamaIndex**: Flexible orchestration and retrieval for agentic workflows.
+*   🌐 **Multilingual Support**: Automatically detects and responds in **English, Spanish, Hindi, and French**.
+*   👤 **Role-Based Adaptation**:
+    *   **Patient Mode**: Simplified, easy-to-understand health information.
+    *   **Doctor Mode**: In-depth, technical insights using specialized medical databases.
+*   🛡️ **Critical Safety Layer**: Identifies keywords related to medical emergencies (e.g., "chest pain") and immediately instructs the user to seek professional help, bypassing the AI.
+*   🔗 **Sourced & Verifiable**: All information is backed by citations from authoritative sources.
+*   💬 **Stateful Conversations**: Remembers chat history for contextual follow-up questions.
+*   🛠️ **Fully Compliant MCP Tool Server**: Our tools can be called externally by any MCP client.
 
 ---
 
-## 🚀 Key Features
-
-* 🌐 **Multilingual Support**: Communicate seamlessly! MediAgent AI automatically detects and responds in **English, Spanish, Hindi, and French**, making it accessible to a global audience.
-* 💬 **Conversational & Stateful**: Enjoy natural, ongoing conversations. The AI remembers the chat history to understand follow-up questions and maintain context throughout your dialogue.
-* 👤 **Role-Based Adaptation**: Get tailored information based on your needs.
-    * **Patient Mode**: Provides simplified, easy-to-understand health advice and general information.
-    * **Doctor Mode**: Delivers in-depth, technical insights, leveraging specialized medical databases for professionals.
-* 🛠️ **Comprehensive Tooling**: MediAgent AI utilizes a robust toolkit to provide accurate answers:
-    * **General Web Search**: For broad health queries.
-    * **PubMed Integration**: Accesses peer-reviewed medical literature for evidence-based information.
-    * **Drug Information**: Provides details on:
-        * **Drug Side Effects**: Find known side effects of medications.
-        * **Drug Interactions**: Check for potential adverse interactions between multiple drugs.
-        * **Official Drug Label Information**: Retrieve detailed FDA-approved prescribing information.
-    * **Clinical Trials Search**: Discover ongoing or past clinical studies relevant to specific conditions or treatments.
-* 🛡️ **Safety First**: Your well-being is paramount. The AI includes a critical safety layer to detect potential medical emergencies and provides appropriate disclaimers, advising users to consult with healthcare professionals for diagnosis and treatment.
-*   **🛠️ MCP Compliant Tool Server**: Our tools can be called externally. **You can test it right now:**
-
-    > ⚠️ **Note:** The first call may take 30-60 seconds as the backend spins up. Subsequent calls are much faster.
-
-    ```bash
-    curl -X POST 'https://sunilpanda992--medi-agent-ai-final-backend-mediagentai-m-3f699b.modal.run' \
-    --header 'Content-Type: application/json' \
-    --data '{
-        "tool_name": "check_drug_interactions",
-        "tool_args": {
-          "drugs": ["warfarin", "amiodarone"]
-        }
-      }'
-    ```
-    
-    **More Example Tool Calls:**
-    
-    - **Drug Side Effects:**
-      ```bash
-      curl -X POST 'https://sunilpanda992--medi-agent-ai-final-backend-mediagentai-m-3f699b.modal.run' \
-      --header 'Content-Type: application/json' \
-      --data '{
-          "tool_name": "check_drug_side_effects",
-          "tool_args": {"drug": "metformin"}
-      }'
-      ```
-    - **PubMed Search:**
-      ```bash
-      curl -X POST 'https://sunilpanda992--medi-agent-ai-final-backend-mediagentai-m-3f699b.modal.run' \
-      --header 'Content-Type: application/json' \
-      --data '{
-          "tool_name": "search_pubmed",
-          "tool_args": {"query": "COVID-19 vaccine efficacy", "max_results": 2}
-      }'
-      ```
-    - **Clinical Trials:**
-      ```bash
-      curl -X POST 'https://sunilpanda992--medi-agent-ai-final-backend-mediagentai-m-3f699b.modal.run' \
-      --header 'Content-Type: application/json' \
-      --data '{
-          "tool_name": "search_clinical_trials",
-          "tool_args": {"query": "diabetes", "max_results": 1}
-      }'
-      ```
-    - **Drug Interaction Warning:**
-      ```bash
-      curl -X POST 'https://sunilpanda992--medi-agent-ai-final-backend-mediagentai-m-3f699b.modal.run' \
-      --header 'Content-Type: application/json' \
-      --data '{
-          "tool_name": "check_drug_interactions",
-          "tool_args": {
-            "drugs": ["aspirin", "ibuprofen"]
-          }
-        }'
-      ```
-    - **Comprehensive Drug Information:**
-      ```bash
-      curl -X POST 'https://sunilpanda992--medi-agent-ai-final-backend-mediagentai-m-3f699b.modal.run' \
-      --header 'Content-Type: application/json' \
-      --data '{
-          "tool_name": "get_drug_info",
-          "tool_args": {"drug": "atorvastatin"}
-      }'
-      ```
-    - **Advanced PubMed Search:**
-      ```bash
-      curl -X POST 'https://sunilpanda992--medi-agent-ai-final-backend-mediagentai-m-3f699b.modal.run' \
-      --header 'Content-Type: application/json' \
-      --data '{
-          "tool_name": "search_pubmed",
-          "tool_args": {"query": "metformin AND cardiovascular outcomes", "max_results": 3}
-      }'
-      ```
-    - **Clinical Trials with Filters:**
-      ```bash
-      curl -X POST 'https://sunilpanda992--medi-agent-ai-final-backend-mediagentai-m-3f699b.modal.run' \
-      --header 'Content-Type: application/json' \
-      --data '{
-          "tool_name": "search_clinical_trials",
-          "tool_args": {"query": "hypertension", "max_results": 2, "filters": {"status": "ongoing"}}
-      }'
-      ```
+## 🛠️ Tools & Technology Stack
 
----
+Our agent leverages a powerful, specialized toolkit to provide accurate information. These tools are exposed via our public MCP endpoint.
 
-## 🔗 How to Use MediAgent AI as an MCP Tool
+| Tool Name                     | Description                                                                 | Data Source(s)             |
+| ----------------------------- | --------------------------------------------------------------------------- | -------------------------- |
+| `search_web`                  | For general health queries and patient-level explanations.                  | DuckDuckGo                 |
+| `search_pubmed`               | Accesses peer-reviewed medical literature for evidence-based information.   | PubMed / NCBI              |
+| `get_drug_label_info`         | Retrieves detailed, official FDA-approved prescribing information.          | openFDA / DailyMed         |
+| `check_drug_side_effects`     | Finds known side effects for a specific medication.                         | openFDA / DailyMed         |
+| `check_drug_interactions`     | Checks for adverse interactions between two or more drugs.                  | RxNorm API (NIH)           |
+| `search_clinical_trials`      | Discovers ongoing or completed clinical studies for conditions/treatments.  | ClinicalTrials.gov (NIH)   |
 
-- **From any MCP-compatible client** (e.g., Claude Desktop, Cursor, Tiny Agents, or another Gradio app), simply point to our `/mcp_tool_endpoint` and use the tool names and arguments as shown above.
-- **See our demo video for a live MCP integration!**
+**Core Technologies:**
+*   **Backend & Compute:** [**Modal**](https://modal.com/) for scalable, serverless GPU and CPU functions.
+*   **AI Agent Framework:** [**LlamaIndex**](https://www.llamaindex.ai/) for orchestrating the agent's think-plan-act loop.
+*   **LLM:** [**Mistral 7B**](https://mistral.ai/) for high-quality reasoning.
+*   **Frontend UI:** [**Gradio**](https://www.gradio.app/) for the interactive chat interface.
 
 ---
 
-## 🧪 Try It Out! (Examples)
-
-Copy and paste one of the examples below into the chat interface to test MediAgent AI's capabilities. Remember to select the appropriate **"patient"** or **"doctor"** role in the UI for tailored responses.
+## 🏛️ Architecture: Hybrid Single-Server MCP
 
-### General Health Queries
+Our single Modal application serves as **both the core Agent and its own MCP-compliant Tool Server**.
 
-* **English**: "What are the most common side effects of metformin?"
-* **Spanish**: "¿Cuáles son los síntomas de la diabetes tipo 2?"
-* **French**: "Comment savoir si j'ai la grippe ou un simple rhume ?"
-* **Hindi**: "उच्च रक्तचाप के लिए घरेलू उपचार क्या हैं?" (What are the home remedies for high blood pressure?)
-
-### Drug-Specific Queries
-
-* "Can I take ibuprofen with lisinopril? (Check for interactions)"
-* "What are the known side effects of amoxicillin?"
-* "Give me the FDA label information for Tylenol."
-
-### Research & Clinical Trials (Best in Doctor Mode)
-
-* "Are there any ongoing clinical trials for Alzheimer's disease?"
-* "What is the latest PubMed literature on efficacy of novel biologics for Crohn's disease?"
+1.  **Agent Endpoint (`/`):** A private endpoint that takes a user query and orchestrates the full agentic workflow (planning, tool selection, synthesis).
+2.  **MCP Tool Endpoint (hidden in backend):** A public endpoint that exposes our individual tools. This makes our application a valid MCP Server, proving compliance and allowing any external MCP client to use our medical tools.
 
 ---
 
-## 🧑‍⚕️ See The Role-Based Difference (Detailed Examples)
+## 🚀 Test our MCP Tool Server Live!
+
+You can call our tools from any MCP-compatible client or directly with `curl`.
+
+> ⚠️ **Note:** The first call may take 30-60 seconds as the serverless backend spins up. Subsequent calls are much faster.
+
+```bash
+# Example: Check for a critical drug interaction
+curl -X POST 'https://sunilpanda992--medi-agent-ai-final-backend-mediagentai-m-3f699b.modal.run' \
+--header 'Content-Type: application/json' \
+--data '{
+    "tool_name": "check_drug_interactions",
+    "tool_args": {
+      "drugs": ["warfarin", "amiodarone"]
+    }
+  }'
+
+# Example: Search for PubMed articles
+curl -X POST 'https://sunilpanda992--medi-agent-ai-final-backend-mediagentai-m-3f699b.modal.run' \
+--header 'Content-Type: application/json' \
+--data '{
+    "tool_name": "search_pubmed",
+    "tool_args": {"query": "metformin AND cardiovascular outcomes", "max_results": 2}
+}'
+
+# Example: Get official FDA label information (CORRECTED)
+curl -X POST 'https://sunilpanda992--medi-agent-ai-final-backend-mediagentai-m-3f699b.modal.run' \
+--header 'Content-Type: application/json' \
+--data '{
+    "tool_name": "get_drug_label_info",
+    "tool_args": {"drug_name": "atorvastatin"}
+}'
+```
+
+IGNORE_WHEN_COPYING_START
+Use code with caution. Markdown
+IGNORE_WHEN_COPYING_END
 
-Select a role in the UI and use a matching prompt below to see how the AI adapts its response, using different tools and language.
-
-### For the Patient Role
-
-Select **patient** and ask a general health question. The AI will primarily use web search to provide an easy-to-understand overview, avoiding overly technical jargon.
+---
+## 🧪 Try It Out in the App!
 
-* "What are the most common side effects I should watch out for with lisinopril?"
+Select the appropriate "patient" or "doctor" role in the chat interface above and try these prompts:
 
-### For the Doctor Role
+    Patient Query: "What are the most common side effects of metformin?"
 
-Select **doctor** and ask a technical, multi-part question. The AI will leverage specialized tools like PubMed, ClinicalTrials.gov, and drug databases to provide a detailed, sourced answer suitable for a medical professional.
+    Doctor Query: "My patient with Crohn's is on methotrexate and we are considering adding infliximab. I need the latest PubMed literature on efficacy and a check for interactions between the two drugs."
 
-* "My patient with Crohn's is on methotrexate and we are considering adding infliximab. I need the latest PubMed literature on efficacy, a check for interactions between the two drugs, and a list of ongoing clinical trials for novel biologics in IBD."
+    Spanish Query: "¿Cuáles son los síntomas de la diabetes tipo 2?"
 
----
+    Interaction Check: "Can I take ibuprofen with lisinopril?"

app.py

@@ -1,3 +1,5 @@
+# In app.py
+
 import gradio as gr
 import requests
 import os
@@ -30,49 +32,38 @@ CUSTOM_CSS = """
 }
 """
 
-def handle_user_turn(user_input, chat_history, user_role, sid):
+def handle_user_turn(user_input: str, chat_history: list, user_role: str, sid: str):
+    """
+    Handles a single turn of the conversation.
+    """
     if not user_input.strip():
         return chat_history, ""
 
-    # Gradio's 'messages' type for chatbot expects dictionaries
-    # Ensure chat_history is initialized with dictionaries or converted
-    if not chat_history or not isinstance(chat_history[0], dict):
-        # Convert existing tuples to dictionaries if needed, or start fresh
-        new_chat_history = []
-        for user_msg, bot_msg in chat_history:
-            if user_msg is not None:
-                new_chat_history.append({"role": "user", "content": user_msg})
-            if bot_msg is not None:
-                new_chat_history.append({"role": "assistant", "content": bot_msg})
-        chat_history = new_chat_history
+    # Initialize chat_history if it's the first turn (it will be None)
+    if chat_history is None:
+        chat_history = []
+    
+    # Append the user's message in the correct dictionary format
+    chat_history.append({"role": "user", "content": user_input})
     
-    chat_history.append({"role": "user", "content": user_input}) # Changed to dictionary format
+    # Immediately yield the updated history to show the user's message
+    yield chat_history, ""
 
     if not MODAL_BACKEND_URL:
-        # For an error, append as assistant message
-        chat_history.append({"role": "assistant", "content": "❗ Configuration error: `MODAL_BACKEND_URL` not set."})
-        return chat_history, ""
+        chat_history.append({"role": "assistant", "content": "❗ Configuration error: `MODAL_BACKEND_URL` is not set."})
+        yield chat_history, ""
+        return
 
-    thinking_steps = ["Analyzing...", "Searching sources...", "Synthesizing..."]
+    # --- Streaming "Thinking" Animation ---
+    chat_history.append({"role": "assistant", "content": ""}) # Add an empty placeholder for the bot response
     
-    # Store the last user message to update it later
-    # This assumes the last message is always the user's latest input
-    last_user_message_index = len(chat_history) - 1
-
+    thinking_steps = ["Analyzing your query...", "Searching medical databases...", "Consulting research papers...", "Synthesizing the answer..."]
     for step in thinking_steps:
-        # Create a temporary 'thinking' message from the assistant
-        temp_message = {"role": "assistant", "content": f"<div class='chatbot-message bot'><em>{step}</em></div>"}
-        
-        # If there's an existing assistant thinking message, update it in place
-        # Otherwise, append a new one
-        if len(chat_history) > last_user_message_index + 1 and chat_history[last_user_message_index + 1]["role"] == "assistant":
-            chat_history[last_user_message_index + 1] = temp_message
-        else:
-            chat_history.append(temp_message)
-
-        yield chat_history, "" # Yield with the temporary thinking message
-
+        # Update the last message (the bot's placeholder) with the current thinking step
+        chat_history[-1]["content"] = f"<em>{step}</em>"
+        yield chat_history, ""
         time.sleep(0.4)
+    # --- End of Animation ---
 
     try:
         response = requests.post(MODAL_BACKEND_URL, json={
@@ -81,38 +72,50 @@ def handle_user_turn(user_input, chat_history, user_role, sid):
             "role": user_role
         }, timeout=600)
         response.raise_for_status()
-        reply = response.json().get("final_markdown", "No response received.")
+        reply = response.json().get("final_markdown", "No valid response was received from the backend.")
+    except requests.exceptions.RequestException as e:
+        reply = f"❗ **Error Communicating with Backend:**\n\nCould not get a response. The server may be starting up or experiencing an issue. Please try again in a moment. \n\n*Details: {str(e)}*"
     except Exception as e:
-        reply = f"❗ Error: {str(e)}"
+        reply = f"❗ **An Unexpected Error Occurred:**\n\n*Details: {str(e)}*"
     
-    # Update the last assistant message with the final reply
-    # This should replace the thinking message or be appended if no thinking message was added
-    if len(chat_history) > last_user_message_index + 1 and chat_history[last_user_message_index + 1]["role"] == "assistant":
-        chat_history[last_user_message_index + 1] = {"role": "assistant", "content": reply}
-    else:
-        chat_history.append({"role": "assistant", "content": reply}) # Should not typically happen if thinking steps run
+    # Update the final bot message with the actual reply
+    chat_history[-1]["content"] = reply
 
     yield chat_history, ""
 
-with gr.Blocks(css=CUSTOM_CSS, title="MediAgent AI") as demo:
+
+# --- The rest of your Gradio app code remains the same ---
+with gr.Blocks(css=CUSTOM_CSS, title="MediAgent AI: Conversational Health Navigator & MCP Tool Server") as demo:
     session_id = gr.State(lambda: str(uuid.uuid4()))
 
     gr.Markdown("""
-    # 🧬 MediAgent AI
-    **Your AI Health Navigator**
-
+    # 🧬 MediAgent AI: Conversational Health Navigator & MCP Tool Server
+    
     <div class='chatbot-box'>
-    MediAgent AI is a multilingual, role-based assistant that gives clear, cited medical answers.
+    <b>MediAgent AI</b> is a multilingual, role-adaptive health assistant and a fully MCP-compliant tool server.<br>
     <ul>
-      <li>🗣 Ask in English, Spanish, Hindi, or French</li>
-      <li>🎯 <b>Patient</b> = simplified advice | <b>Doctor</b> = in-depth, technical insights</li>
-      <li>🔗 Answers are backed by PubMed, FDA, etc.</li>
-      <li>💊 Can provide information on **drug side effects, interactions, and official label details.**</li>
-      <li>🔬 Can search for **clinical trials and medical research.**</li>
-      <li>🛡️ Detects emergencies, adds safety notices</li>
+      <li>🌐 <b>Multilingual:</b> English, Spanish, Hindi, French</li>
+      <li>👤 <b>Role-Based:</b> <b>Patient</b> = simplified advice | <b>Doctor</b> = in-depth, technical insights</li>
+      <li>🛡️ <b>Critical Safety Layer:</b> Detects emergencies, adds safety notices</li>
+      <li>🔗 <b>Sourced & Verifiable:</b> Answers are backed by PubMed, FDA, etc.</li>
+      <li>💬 <b>Stateful Conversations:</b> Remembers chat history for contextual follow-up</li>
+      <li>🛠️ <b>MCP Tool Server:</b> All tools are externally callable via MCP</li>
+      <li>💊 <b>Drug Info:</b> Side effects, interactions, FDA label details</li>
+      <li>🔬 <b>Research:</b> Clinical trials & medical literature search</li>
     </ul>
+    <div style='margin-top:10px; text-align:center;'>
+      <a href='https://modal.com/' target='_blank' style='text-decoration:none; margin-right:18px;'>
+        <img src='https://assets.modal.com/logos/modal-logo-dark.svg' alt='Modal Logo' style='height:32px; vertical-align:middle; margin-right:8px;'>
+        <span style='font-size:1.1em; color:#8aa8ff; font-weight:bold;'>Powered by Modal</span>
+      </a>
+      <a href='https://www.llamaindex.ai/' target='_blank' style='text-decoration:none;'>
+        <img src='https://raw.githubusercontent.com/jerryjliu/llama_index/main/docs/imgs/llama_index_logo.png' alt='LlamaIndex Logo' style='height:32px; vertical-align:middle; margin-right:8px;'>
+        <span style='font-size:1.1em; color:#ffb300; font-weight:bold;'>and LlamaIndex</span>
+      </a>
+      <div style='font-size:0.95em; color:#aaa; margin-top:2px;'>Special thanks to the Modal and LlamaIndex teams for their amazing platforms!</div>
     </div>
-    """) # Changed Markdown content
+    </div>
+    """)
 
     with gr.Row():
         with gr.Column(scale=1):
@@ -123,7 +126,8 @@ with gr.Blocks(css=CUSTOM_CSS, title="MediAgent AI") as demo:
                 info="Choose your user type for tailored responses"
             )
         with gr.Column(scale=3):
-            chatbot = gr.Chatbot(label="Conversation", height=480, type='messages') # Added type='messages'
+            # IMPORTANT: Ensure chatbot is defined with type='messages'
+            chatbot = gr.Chatbot(label="Conversation", height=480, type='messages', bubble_full_width=False)
 
     with gr.Row():
         user_textbox = gr.Textbox(
@@ -131,34 +135,35 @@ with gr.Blocks(css=CUSTOM_CSS, title="MediAgent AI") as demo:
             lines=1,
             scale=4
         )
-        submit_btn = gr.Button("Send", elem_id="send-button", scale=1)
+        submit_btn = gr.Button("Send", elem_id="send-button", scale=1, variant="primary")
 
     examples = gr.Examples(
         examples=[
             "What are the side effects of metformin?",
             "Explain Crohn’s disease vs Ulcerative Colitis for a doctor",
             "¿Cuáles son los síntomas de la hipertensión?",
-            "Can I take ibuprofen with lisinopril? (for interactions)", # New example
-            "What is the FDA label information for acetaminophen?", # New example
-            "Are there any ongoing clinical trials for breast cancer?", # New example
+            "Can I take ibuprofen with lisinopril? (for interactions)",
+            "What is the FDA label information for acetaminophen?",
+            "Are there any ongoing clinical trials for breast cancer?",
         ],
         inputs=[user_textbox]
-    ) # Changed Examples content
+    )
 
-    gr.ClearButton([chatbot, user_textbox], value="🔄 Clear")
+    gr.ClearButton([chatbot, user_textbox], value="🔄 Clear Chat")
 
-    submit_btn.click(
+    # Connect submit actions
+    user_textbox.submit(
         fn=handle_user_turn,
         inputs=[user_textbox, chatbot, user_role, session_id],
         outputs=[chatbot, user_textbox]
     )
-
-    user_textbox.submit(
+    submit_btn.click(
         fn=handle_user_turn,
         inputs=[user_textbox, chatbot, user_role, session_id],
         outputs=[chatbot, user_textbox]
     )
 
+
 if __name__ == "__main__":
     demo.queue()
     demo.launch()