PraisonAI 0.0.73__tar.gz → 0.0.74__tar.gz

{praisonai-0.0.73 → praisonai-0.0.74}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: PraisonAI
-Version: 0.0.73
+Version: 0.0.74
 Summary: PraisonAI application combines AutoGen and CrewAI or similar frameworks into a low-code solution for building and managing multi-agent LLM systems, focusing on simplicity, customization, and efficient human-agent collaboration.
 Author: Mervin Praison
 Requires-Python: >=3.10,<3.13
@@ -17,27 +17,30 @@ Provides-Extra: cohere
 Provides-Extra: google
 Provides-Extra: gradio
 Provides-Extra: openai
+Provides-Extra: realtime
 Provides-Extra: train
 Provides-Extra: ui
 Requires-Dist: agentops (>=0.3.12) ; extra == "agentops"
-Requires-Dist: aiosqlite (>=0.20.0) ; extra == "chat" or extra == "code"
-Requires-Dist: chainlit (==1.2.0) ; extra == "ui" or extra == "chat" or extra == "code"
-Requires-Dist: crawl4ai (==0.3.4) ; extra == "chat" or extra == "code"
+Requires-Dist: aiosqlite (>=0.20.0) ; extra == "chat" or extra == "code" or extra == "realtime"
+Requires-Dist: chainlit (==1.3.0rc1) ; extra == "ui" or extra == "chat" or extra == "code" or extra == "realtime"
+Requires-Dist: crawl4ai (==0.3.4) ; extra == "chat" or extra == "code" or extra == "realtime"
 Requires-Dist: crewai (>=0.32.0)
 Requires-Dist: flask (>=3.0.0) ; extra == "api"
 Requires-Dist: gradio (>=4.26.0) ; extra == "gradio"
-Requires-Dist: greenlet (>=3.0.3) ; extra == "chat" or extra == "code"
+Requires-Dist: greenlet (>=3.0.3) ; extra == "chat" or extra == "code" or extra == "realtime"
 Requires-Dist: langchain-anthropic (>=0.1.13) ; extra == "anthropic"
 Requires-Dist: langchain-cohere (>=0.1.4) ; extra == "cohere"
 Requires-Dist: langchain-google-genai (>=1.0.4) ; extra == "google"
 Requires-Dist: langchain-openai (>=0.1.7) ; extra == "openai"
-Requires-Dist: litellm (>=1.41.8) ; extra == "chat" or extra == "code"
+Requires-Dist: litellm (>=1.41.8) ; extra == "chat" or extra == "code" or extra == "realtime"
 Requires-Dist: markdown (>=3.5)
+Requires-Dist: plotly (>=5.24.0) ; extra == "realtime"
 Requires-Dist: praisonai-tools (>=0.0.7)
 Requires-Dist: pyautogen (>=0.2.19)
 Requires-Dist: pyparsing (>=3.0.0)
 Requires-Dist: rich (>=13.7)
-Requires-Dist: tavily-python (==0.5.0) ; extra == "chat" or extra == "code"
+Requires-Dist: tavily-python (==0.5.0) ; extra == "chat" or extra == "code" or extra == "realtime"
+Requires-Dist: websockets (>=12.0) ; extra == "realtime"
 Project-URL: Homepage, https://docs.praison.ai
 Project-URL: Repository, https://github.com/mervinpraison/PraisonAI
 Description-Content-Type: text/markdown
@@ -79,6 +82,7 @@ Praison AI, leveraging both AutoGen and CrewAI or any other agent framework, rep
 | **UI** | Multi Agents such as CrewAI or AutoGen | [https://docs.praison.ai/ui/ui](https://docs.praison.ai/ui/ui) |
 | **Chat** | Chat with 100+ LLMs, single AI Agent | [https://docs.praison.ai/ui/chat](https://docs.praison.ai/ui/chat) |
 | **Code** | Chat with entire Codebase, single AI Agent | [https://docs.praison.ai/ui/code](https://docs.praison.ai/ui/code) |
+| **Realtime** | Real-time voice interaction with AI | [https://docs.praison.ai/ui/realtime](https://docs.praison.ai/ui/realtime) |
 
 | Other Features | Description | Docs |
 |---|---|---|
@@ -100,6 +104,7 @@ Praison AI, leveraging both AutoGen and CrewAI or any other agent framework, rep
 | **PraisonAI Code** | `pip install "praisonai[code]"` |
 | **PraisonAI Chat** | `pip install "praisonai[chat]"` |
 | **PraisonAI Train** | `pip install "praisonai[train]"` |
+| **PraisonAI Realtime** | `pip install "praisonai[realtime]"` |
 
 ## Key Features
 
@@ -110,6 +115,9 @@ Praison AI, leveraging both AutoGen and CrewAI or any other agent framework, rep
 - Interactive UIs
 - YAML-based Configuration
 - Custom Tool Integration
+- Internet Search Capability (using Crawl4AI and Tavily)
+- Vision Language Model (VLM) Support
+- Real-time Voice Interaction
 
 ## TL;DR Multi Agents
 
@@ -222,6 +230,14 @@ export OPENAI_API_KEY="Enter your API key"
 praisonai chat
 ```
 
+### Internet Search
+
+Praison AI Chat and Praison AI Code now includes internet search capabilities using Crawl4AI and Tavily, allowing you to retrieve up-to-date information during your conversations.
+
+### Vision Language Model Support
+
+You can now upload images and ask questions based on them using Vision Language Models. This feature enables visual understanding and analysis within your chat sessions.
+
 ## Praison AI Code
 
 ```bash
@@ -230,6 +246,10 @@ export OPENAI_API_KEY="Enter your API key"
 praisonai code
 ```
 
+### Internet Search
+
+Praison AI Code also includes internet search functionality, enabling you to find relevant code snippets and programming information online.
+
 ## Create Custom Tools
 
 - https://docs.praison.ai/tools/custom/

{praisonai-0.0.73 → praisonai-0.0.74}/README.md

@@ -35,6 +35,7 @@ Praison AI, leveraging both AutoGen and CrewAI or any other agent framework, rep
 | **UI** | Multi Agents such as CrewAI or AutoGen | [https://docs.praison.ai/ui/ui](https://docs.praison.ai/ui/ui) |
 | **Chat** | Chat with 100+ LLMs, single AI Agent | [https://docs.praison.ai/ui/chat](https://docs.praison.ai/ui/chat) |
 | **Code** | Chat with entire Codebase, single AI Agent | [https://docs.praison.ai/ui/code](https://docs.praison.ai/ui/code) |
+| **Realtime** | Real-time voice interaction with AI | [https://docs.praison.ai/ui/realtime](https://docs.praison.ai/ui/realtime) |
 
 | Other Features | Description | Docs |
 |---|---|---|
@@ -56,6 +57,7 @@ Praison AI, leveraging both AutoGen and CrewAI or any other agent framework, rep
 | **PraisonAI Code** | `pip install "praisonai[code]"` |
 | **PraisonAI Chat** | `pip install "praisonai[chat]"` |
 | **PraisonAI Train** | `pip install "praisonai[train]"` |
+| **PraisonAI Realtime** | `pip install "praisonai[realtime]"` |
 
 ## Key Features
 
@@ -66,6 +68,9 @@ Praison AI, leveraging both AutoGen and CrewAI or any other agent framework, rep
 - Interactive UIs
 - YAML-based Configuration
 - Custom Tool Integration
+- Internet Search Capability (using Crawl4AI and Tavily)
+- Vision Language Model (VLM) Support
+- Real-time Voice Interaction
 
 ## TL;DR Multi Agents
 
@@ -178,6 +183,14 @@ export OPENAI_API_KEY="Enter your API key"
 praisonai chat
 ```
 
+### Internet Search
+
+Praison AI Chat and Praison AI Code now includes internet search capabilities using Crawl4AI and Tavily, allowing you to retrieve up-to-date information during your conversations.
+
+### Vision Language Model Support
+
+You can now upload images and ask questions based on them using Vision Language Models. This feature enables visual understanding and analysis within your chat sessions.
+
 ## Praison AI Code
 
 ```bash
@@ -186,6 +199,10 @@ export OPENAI_API_KEY="Enter your API key"
 praisonai code
 ```
 
+### Internet Search
+
+Praison AI Code also includes internet search functionality, enabling you to find relevant code snippets and programming information online.
+
 ## Create Custom Tools
 
 - https://docs.praison.ai/tools/custom/

{praisonai-0.0.73 → praisonai-0.0.74}/praisonai/cli.py

@@ -130,6 +130,10 @@ class PraisonAI:
             self.create_code_interface()
             return
         
+        if getattr(args, 'realtime', False):
+            self.create_realtime_interface()
+            return
+        
         if args.agent_file == 'train':
             package_root = os.path.dirname(os.path.abspath(__file__))
             config_yaml_destination = os.path.join(os.getcwd(), 'config.yaml')
@@ -256,6 +260,7 @@ class PraisonAI:
         parser.add_argument("--hf", type=str, help="Hugging Face model name")
         parser.add_argument("--ollama", type=str, help="Ollama model name")
         parser.add_argument("--dataset", type=str, help="Dataset name for training", default="yahma/alpaca-cleaned")
+        parser.add_argument("--realtime", action="store_true", help="Start the realtime voice interaction interface")
         args, unknown_args = parser.parse_known_args()
 
         if unknown_args and unknown_args[0] == '-b' and unknown_args[1] == 'api:app':
@@ -270,6 +275,8 @@ class PraisonAI:
         if args.agent_file == 'code':
             args.ui = 'chainlit'
             args.code = True
+        if args.agent_file == 'realtime':
+            args.realtime = True
 
         return args
     
@@ -416,6 +423,29 @@ class PraisonAI:
         else:
             print("ERROR: Chainlit is not installed. Please install it with 'pip install \"praisonai\[ui]\"' to use the UI.")        
 
+    def create_realtime_interface(self):
+        """
+        Create a Chainlit interface for the realtime voice interaction application.
+        """
+        if CHAINLIT_AVAILABLE:
+            import praisonai
+            os.environ["CHAINLIT_PORT"] = "8088"  # Ensure this port is not in use by another service
+            root_path = os.path.join(os.path.expanduser("~"), ".praison")
+            os.environ["CHAINLIT_APP_ROOT"] = root_path
+            public_folder = os.path.join(os.path.dirname(praisonai.__file__), 'public')
+            if not os.path.exists(os.path.join(root_path, "public")):
+                if os.path.exists(public_folder):
+                    shutil.copytree(public_folder, os.path.join(root_path, "public"), dirs_exist_ok=True)
+                    logging.info("Public folder copied successfully!")
+                else:
+                    logging.info("Public folder not found in the package.")
+            else:
+                logging.info("Public folder already exists.")
+            realtime_ui_path = os.path.join(os.path.dirname(praisonai.__file__), 'ui', 'realtime.py')
+            chainlit_run([realtime_ui_path])
+        else:
+            print("ERROR: Realtime UI is not installed. Please install it with 'pip install \"praisonai[realtime]\"' to use the realtime UI.")
+
 if __name__ == "__main__":
     praison_ai = PraisonAI()
     praison_ai.main()

{praisonai-0.0.73 → praisonai-0.0.74}/praisonai/deploy.py

@@ -56,7 +56,7 @@ class CloudDeployer:
             file.write("FROM python:3.11-slim\n")
             file.write("WORKDIR /app\n")
             file.write("COPY . .\n")
-            file.write("RUN pip install flask praisonai==0.0.73 gunicorn markdown\n")
+            file.write("RUN pip install flask praisonai==0.0.74 gunicorn markdown\n")
             file.write("EXPOSE 8080\n")
             file.write('CMD ["gunicorn", "-b", "0.0.0.0:8080", "api:app"]\n')
             

praisonai-0.0.74/praisonai/ui/realtime.py

@@ -0,0 +1,368 @@
+import os
+import asyncio
+import sqlite3
+from datetime import datetime
+from uuid import uuid4
+
+from openai import AsyncOpenAI
+import chainlit as cl
+from chainlit.logger import logger
+from chainlit.input_widget import TextInput
+from chainlit.types import ThreadDict
+
+from realtimeclient import RealtimeClient
+from realtimeclient.tools import tools
+from sql_alchemy import SQLAlchemyDataLayer
+import chainlit.data as cl_data
+from literalai.helper import utc_now
+import json
+
+# Set up CHAINLIT_AUTH_SECRET
+CHAINLIT_AUTH_SECRET = os.getenv("CHAINLIT_AUTH_SECRET")
+
+if not CHAINLIT_AUTH_SECRET:
+    os.environ["CHAINLIT_AUTH_SECRET"] = "p8BPhQChpg@J>jBz$wGxqLX2V>yTVgP*7Ky9H$aV:axW~ANNX-7_T:o@lnyCBu^U"
+    CHAINLIT_AUTH_SECRET = os.getenv("CHAINLIT_AUTH_SECRET")
+
+# Database path
+DB_PATH = os.path.expanduser("~/.praison/database.sqlite")
+
+def initialize_db():
+    os.makedirs(os.path.dirname(DB_PATH), exist_ok=True)
+    conn = sqlite3.connect(DB_PATH)
+    cursor = conn.cursor()
+    cursor.execute('''
+        CREATE TABLE IF NOT EXISTS users (
+            id UUID PRIMARY KEY,
+            identifier TEXT NOT NULL UNIQUE,
+            metadata JSONB NOT NULL,
+            createdAt TEXT
+        )
+    ''')
+    cursor.execute('''
+        CREATE TABLE IF NOT EXISTS threads (
+            id UUID PRIMARY KEY,
+            createdAt TEXT,
+            name TEXT,
+            userId UUID,
+            userIdentifier TEXT,
+            tags TEXT[],
+            metadata JSONB NOT NULL DEFAULT '{}',
+            FOREIGN KEY (userId) REFERENCES users(id) ON DELETE CASCADE
+        )
+    ''')
+    cursor.execute('''
+        CREATE TABLE IF NOT EXISTS steps (
+            id UUID PRIMARY KEY,
+            name TEXT NOT NULL,
+            type TEXT NOT NULL,
+            threadId UUID NOT NULL,
+            parentId UUID,
+            disableFeedback BOOLEAN NOT NULL DEFAULT 0,
+            streaming BOOLEAN NOT NULL DEFAULT 0,
+            waitForAnswer BOOLEAN DEFAULT 0,
+            isError BOOLEAN NOT NULL DEFAULT 0,
+            metadata JSONB DEFAULT '{}',
+            tags TEXT[],
+            input TEXT,
+            output TEXT,
+            createdAt TEXT,
+            start TEXT,
+            end TEXT,
+            generation JSONB,
+            showInput TEXT,
+            language TEXT,
+            indent INT,
+            FOREIGN KEY (threadId) REFERENCES threads (id) ON DELETE CASCADE
+        )
+    ''')
+    cursor.execute('''
+        CREATE TABLE IF NOT EXISTS elements (
+            id UUID PRIMARY KEY,
+            threadId UUID,
+            type TEXT,
+            url TEXT,
+            chainlitKey TEXT,
+            name TEXT NOT NULL,
+            display TEXT,
+            objectKey TEXT,
+            size TEXT,
+            page INT,
+            language TEXT,
+            forId UUID,
+            mime TEXT,
+            FOREIGN KEY (threadId) REFERENCES threads (id) ON DELETE CASCADE
+        )
+    ''')
+    cursor.execute('''
+        CREATE TABLE IF NOT EXISTS feedbacks (
+            id UUID PRIMARY KEY,
+            forId UUID NOT NULL,
+            value INT NOT NULL,
+            threadId UUID,
+            comment TEXT
+        )
+    ''')
+    cursor.execute('''
+        CREATE TABLE IF NOT EXISTS settings (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            key TEXT UNIQUE,
+            value TEXT
+        )
+    ''')
+    conn.commit()
+    conn.close()
+
+def save_setting(key: str, value: str):
+    """Saves a setting to the database."""
+    conn = sqlite3.connect(DB_PATH)
+    cursor = conn.cursor()
+    cursor.execute(
+        """
+        INSERT OR REPLACE INTO settings (id, key, value)
+        VALUES ((SELECT id FROM settings WHERE key = ?), ?, ?)
+    """,
+        (key, key, value),
+    )
+    conn.commit()
+    conn.close()
+
+def load_setting(key: str) -> str:
+    """Loads a setting from the database."""
+    conn = sqlite3.connect(DB_PATH)
+    cursor = conn.cursor()
+    cursor.execute('SELECT value FROM settings WHERE key = ?', (key,))
+    result = cursor.fetchone()
+    conn.close()
+    return result[0] if result else None
+
+# Initialize the database
+initialize_db()
+
+# Set up SQLAlchemy data layer
+cl_data._data_layer = SQLAlchemyDataLayer(conninfo=f"sqlite+aiosqlite:///{DB_PATH}")
+
+client = AsyncOpenAI()
+
+@cl.on_chat_start
+async def start():
+    initialize_db()
+    model_name = load_setting("model_name") or os.getenv("MODEL_NAME", "gpt-4o-mini")
+    cl.user_session.set("model_name", model_name)
+    cl.user_session.set("message_history", [])  # Initialize message history
+    logger.debug(f"Model name: {model_name}")
+    # settings = cl.ChatSettings(
+    #     [
+    #         TextInput(
+    #             id="model_name",
+    #             label="Enter the Model Name",
+    #             placeholder="e.g., gpt-4o-mini",
+    #             initial=model_name
+    #         )
+    #     ]
+    # )
+    # cl.user_session.set("settings", settings)
+    # await settings.send()
+    await cl.Message(
+        content="Welcome to the PraisonAI realtime. Press `P` to talk!"
+    ).send()
+    await setup_openai_realtime()
+
+@cl.on_message
+async def on_message(message: cl.Message):
+    openai_realtime: RealtimeClient = cl.user_session.get("openai_realtime")
+    message_history = cl.user_session.get("message_history", [])
+    
+    if openai_realtime and openai_realtime.is_connected():
+        current_date = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
+        prompt = f"Current time Just for reference: {current_date}\n\n{message.content}"
+        
+        # Add user message to history
+        message_history.append({"role": "user", "content": prompt})
+        cl.user_session.set("message_history", message_history)
+        
+        await openai_realtime.send_user_message_content([{ "type": 'input_text', "text": message.content }])
+    else:
+        await cl.Message(content="Please activate voice mode before sending messages!").send()
+
+async def setup_openai_realtime():
+    """Instantiate and configure the OpenAI Realtime Client"""
+    openai_realtime = RealtimeClient(api_key=os.getenv("OPENAI_API_KEY"))
+    cl.user_session.set("track_id", str(uuid4()))
+
+    async def handle_conversation_updated(event):
+        item = event.get("item")
+        delta = event.get("delta")
+        """Currently used to stream audio back to the client."""
+        if delta:
+            if 'audio' in delta:
+                audio = delta['audio']  # Int16Array, audio added
+                await cl.context.emitter.send_audio_chunk(cl.OutputAudioChunk(mimeType="pcm16", data=audio, track=cl.user_session.get("track_id")))
+            if 'transcript' in delta:
+                transcript = delta['transcript']  # string, transcript added
+                logger.debug(f"Transcript delta: {transcript}")
+            if 'text' in delta:
+                text = delta['text']  # string, text added
+                logger.debug(f"Text delta: {text}")
+            if 'arguments' in delta:
+                arguments = delta['arguments']  # string, function arguments added
+                logger.debug(f"Function arguments delta: {arguments}")
+
+    async def handle_item_completed(event):
+        """Used to populate the chat context with transcription once an item is completed."""
+        try:
+            item = event.get("item")
+            logger.debug(f"Item completed: {json.dumps(item, indent=2, default=str)}")
+            await openai_realtime._send_chainlit_message(item)
+            
+            # Add assistant message to history
+            message_history = cl.user_session.get("message_history", [])
+            content = item.get("formatted", {}).get("text", "") or item.get("formatted", {}).get("transcript", "")
+            if content:
+                message_history.append({"role": "assistant", "content": content})
+                cl.user_session.set("message_history", message_history)
+        except Exception as e:
+            error_message = f"Error in handle_item_completed: {str(e)}"
+            logger.error(error_message)
+            debug_item = json.dumps(item, indent=2, default=str)
+            logger.error(f"Item causing error: {debug_item}")
+
+    async def handle_conversation_interrupt(event):
+        """Used to cancel the client previous audio playback."""
+        cl.user_session.set("track_id", str(uuid4()))
+        await cl.context.emitter.send_audio_interrupt()
+
+    async def handle_error(event):
+        logger.error(event)
+        await cl.Message(content=f"Error: {event}", author="System").send()
+
+    openai_realtime.on('conversation.updated', handle_conversation_updated)
+    openai_realtime.on('conversation.item.completed', handle_item_completed)
+    openai_realtime.on('conversation.interrupted', handle_conversation_interrupt)
+    openai_realtime.on('error', handle_error)
+
+    cl.user_session.set("openai_realtime", openai_realtime)
+    coros = [openai_realtime.add_tool(tool_def, tool_handler) for tool_def, tool_handler in tools]
+    await asyncio.gather(*coros)
+
+@cl.on_settings_update
+async def setup_agent(settings):
+    logger.debug(settings)
+    cl.user_session.set("settings", settings)
+    model_name = settings["model_name"]
+    cl.user_session.set("model_name", model_name)
+    
+    # Save in settings table
+    save_setting("model_name", model_name)
+    
+    # Save in thread metadata
+    thread_id = cl.user_session.get("thread_id")
+    if thread_id:
+        thread = await cl_data._data_layer.get_thread(thread_id)
+        if thread:
+            metadata = thread.get("metadata", {})
+            if isinstance(metadata, str):
+                try:
+                    metadata = json.loads(metadata)
+                except json.JSONDecodeError:
+                    metadata = {}
+            
+            metadata["model_name"] = model_name
+            
+            # Always store metadata as a dictionary
+            await cl_data._data_layer.update_thread(thread_id, metadata=metadata)
+            
+            # Update the user session with the new metadata
+            cl.user_session.set("metadata", metadata)
+
+@cl.on_audio_start
+async def on_audio_start():
+    try:
+        openai_realtime: RealtimeClient = cl.user_session.get("openai_realtime")
+        await openai_realtime.connect()
+        logger.info("Connected to OpenAI realtime")
+        return True
+    except Exception as e:
+        await cl.ErrorMessage(content=f"Failed to connect to OpenAI realtime: {e}").send()
+        return False
+
+@cl.on_audio_chunk
+async def on_audio_chunk(chunk: cl.InputAudioChunk):
+    openai_realtime: RealtimeClient = cl.user_session.get("openai_realtime")            
+    if openai_realtime.is_connected():
+        await openai_realtime.append_input_audio(chunk.data)
+    else:
+        logger.info("RealtimeClient is not connected")
+
+@cl.on_audio_end
+@cl.on_chat_end
+@cl.on_stop
+async def on_end():
+    openai_realtime: RealtimeClient = cl.user_session.get("openai_realtime")
+    if openai_realtime and openai_realtime.is_connected():
+        await openai_realtime.disconnect()
+
+@cl.password_auth_callback
+def auth_callback(username: str, password: str):
+    # You can customize this function to use your own authentication logic
+    expected_username = os.getenv("CHAINLIT_USERNAME", "admin")
+    expected_password = os.getenv("CHAINLIT_PASSWORD", "admin")
+    if (username, password) == (expected_username, expected_password):
+        return cl.User(
+            identifier=username, metadata={"role": "ADMIN", "provider": "credentials"}
+        )
+    else:
+        return None
+
+@cl.on_chat_resume
+async def on_chat_resume(thread: ThreadDict):
+    logger.info(f"Resuming chat: {thread['id']}")
+    model_name = load_setting("model_name") or os.getenv("MODEL_NAME") or "gpt-4o-mini"
+    logger.debug(f"Model name: {model_name}")
+    settings = cl.ChatSettings(
+        [
+            TextInput(
+                id="model_name",
+                label="Enter the Model Name",
+                placeholder="e.g., gpt-4o-mini",
+                initial=model_name
+            )
+        ]
+    )
+    await settings.send()
+    thread_id = thread["id"]
+    cl.user_session.set("thread_id", thread["id"])
+    
+    # Ensure metadata is a dictionary
+    metadata = thread.get("metadata", {})
+    if isinstance(metadata, str):
+        try:
+            metadata = json.loads(metadata)
+        except json.JSONDecodeError:
+            metadata = {}
+    
+    cl.user_session.set("metadata", metadata)
+    
+    message_history = []
+    steps = thread["steps"]
+
+    for message in steps:
+        msg_type = message.get("type")
+        if msg_type == "user_message":
+            message_history.append({"role": "user", "content": message.get("output", "")})
+        elif msg_type == "assistant_message":
+            message_history.append({"role": "assistant", "content": message.get("output", "")})
+        elif msg_type == "run":
+            # Handle 'run' type messages
+            if message.get("isError"):
+                message_history.append({"role": "system", "content": f"Error: {message.get('output', '')}"})
+            else:
+                # You might want to handle non-error 'run' messages differently
+                pass
+        else:
+            logger.warning(f"Message without recognized type: {message}")
+
+    cl.user_session.set("message_history", message_history)
+
+    # Reconnect to OpenAI realtime
+    await setup_openai_realtime()