diff --git a/examples/example.py b/examples/example.py
index 0c9dae76..6214b265 100644
--- a/examples/example.py
+++ b/examples/example.py
@@ -4,6 +4,7 @@
 
 import asyncio
 import sys
+import time
 
 import uvicorn
 from fastapi import FastAPI
@@ -18,17 +19,19 @@
 async def main():
     """Main"""
     talky = MyLLM()
+    # asyncio.ensure_future(async_foo())
 
     logger.info(await talky.chat("My name is Jack"))
     # Hello Jack, it's nice to meet you!
     # I am an AI language model designed to provide helpful responses.
     #  How can I help you today?
+    time.sleep(10)
     logger.info(await talky.chat("tell me who is president of the united states?"))
-    #  As of my latest update, the current
-    # President of the United States is Joe Biden.
-    # He was inaugurated on January 20th, 2021 and
-    # is the 46th President of the United States.
-
+    # #  As of my latest update, the current
+    # # President of the United States is Joe Biden.
+    # # He was inaugurated on January 20th, 2021 and
+    # # is the 46th President of the United States.
+    time.sleep(10)
     logger.info(await talky.chat("what is my name"))
     #  Your name is Jack, as you mentioned earlier.
 
diff --git a/myllm/default_settings.toml b/myllm/default_settings.toml
index 3dd7e646..e11d47e3 100644
--- a/myllm/default_settings.toml
+++ b/myllm/default_settings.toml
@@ -25,8 +25,8 @@ llm_model = "gpt-3.5-turbo"
 # LLM Provider
 # Refer to https://github.com/xtekky/gpt4free
 # for the list of supported provider
-llm_provider = "g4f.Provider.ChatgptAi"
-
+# llm_provider = "g4f.Provider.ChatgptAi"
+llm_provider = "g4f.Provider.Bing"
 # help message listing the commands
 # available
 llm_commands = """
@@ -48,29 +48,32 @@ llm_ai_mode = false
 # template prompt context
 # not implemented
 llm_template = """
-You operate within the following constraints:
-Your task is to devise up to 5 highly effective goals
-and an appropriate role-based name (_GPT) 
-for an autonomous agent, ensuring that the goals 
-are optimally aligned with the successful
-completion of its assigned task.
+test
+"""
+# llm_template = """
+# You operate within the following constraints:
+# Your task is to devise up to 5 highly effective goals
+# and an appropriate role-based name (_GPT) 
+# for an autonomous agent, ensuring that the goals 
+# are optimally aligned with the successful
+# completion of its assigned task.
 
-The user will provide the task, 
-you will provide only the output 
-in the exact format specified below
-with no explanation or conversation.
+# The user will provide the task, 
+# you will provide only the output 
+# in the exact format specified below
+# with no explanation or conversation.
 
-Example input:
-Help me with a trading monitoring for EURUSD
+# Example input:
+# Help me with a trading monitoring for EURUSD
 
-Example output:
-- Query current trend via /trend
-- Access realtime data via /indicator
-- Query news via /news
-- Advise on a low timeframe and high timeframe trading positioning
-- Place an order if suitable and wait for user confirmation
+# Example output:
+# - Query current trend via /trend
+# - Access realtime data via /indicator
+# - Query news via /news
+# - Advise on a low timeframe and high timeframe trading positioning
+# - Place an order if suitable and wait for user confirmation
 
-"""
+# """
 
 ########################################
 ###     END OF DEFAULT SETTINGS      ###
diff --git a/myllm/main.py b/myllm/main.py
index 5b954ef4..be685dd0 100644
--- a/myllm/main.py
+++ b/myllm/main.py
@@ -6,15 +6,10 @@
 
 import asyncio
 import importlib
-from typing import Any, List, Mapping, Optional
 
 import g4f
-import nest_asyncio
+
 from g4f import Provider
-from langchain.chains import ConversationChain
-from langchain.llms.base import LLM
-from langchain.memory import ConversationBufferMemory
-from langchain.prompts import PromptTemplate
 from loguru import logger
 
 from myllm import __version__
@@ -50,14 +45,16 @@ def __init__(self):
             None
         """
 
-        self.logger = logger
         self.enabled = settings.llm_enabled
         if not self.enabled:
             return
         self.commands = settings.llm_commands
         self.llm_ai_mode = settings.llm_ai_mode
-        self.llm = LangLLM()
-        self.conversation = None
+        provider_module_name = settings.llm_provider
+        provider_module = importlib.import_module(provider_module_name)
+        provider_class = getattr(provider_module, provider_module_name.split(".")[-1])
+        self.provider = provider_class()
+        self.conversation = Conversation()
 
     async def get_myllm_info(self):
         """
@@ -89,21 +86,24 @@ async def chat(self, prompt):
         Returns:
             str: The predicted response from the conversation model.
         """
-
-        if self.conversation is None:
-            self.conversation = ConversationChain(
-                llm=self.llm,
-                # prompt=PromptTemplate(template=settings.llm_template),
-                memory=ConversationBufferMemory(),
-            )
-        return self.conversation.predict(input=prompt)
+        logger.debug("chat {}", prompt)
+        self.conversation.add_message("user", prompt)
+        logger.debug("conversation {}", self.conversation.get_messages())
+        response = await self.provider.create_async(
+            model=settings.llm_model,
+            messages=self.conversation.get_messages(),
+        )
+        logger.debug("response {}", response)
+        self.conversation.add_message("ai", response)
+        logger.debug("conversation {}", self.conversation.get_messages())
+        return response
 
     async def clear_chat_history(self):
         """
         Clears the chat history by setting the `conversation`
         attribute to an empty string.
         """
-        self.conversation = ""
+        self.conversation = Conversation()
 
     async def switch_continous_mode(self):
         """ """
@@ -111,45 +111,16 @@ async def switch_continous_mode(self):
         return f"Continous mode {'enabled' if self.llm_ai_mode else 'disabled'}."
 
 
-class LangLLM(LLM):
-    @property
-    def _llm_type(self) -> str:
-        """
-        Returns the type of the _llm_type property.
-
-        :return: A string representing the type of the property.
-        :rtype: str
-        """
-        return "custom"
-
-    def _call(self, prompt: str, stop: Optional[List[str]] = None) -> str:
-        """
-        Calls the ChatCompletion API to generate a response based on the given prompt.
-
-        Args:
-            prompt (str): The prompt for the ChatCompletion API.
-            stop (Optional[List[str]], optional): A list of strings that,
-            if found in the response,
-                indicates the response should be truncated. Defaults to None.
+class Conversation:
+    def __init__(self, max_memory=5):
+        self.messages = []
+        self.max_memory = max_memory
 
-        Returns:
-            str: The generated response from the ChatCompletion API.
-        """
+    def add_message(self, role: str, content: str):
+        if len(self.messages) >= self.max_memory:
+            self.messages.pop(0)  # Remove the oldest message
+        self.messages.append({"role": role, "content": content})
 
-        nest_asyncio.apply()
-        provider_module_name = settings.llm_provider
-        provider_module = importlib.import_module(provider_module_name)
-        provider_class = getattr(provider_module, provider_module_name.split(".")[-1])
-        provider = provider_class()
-
-        out = g4f.ChatCompletion.create(
-            model=settings.llm_model,
-            provider=provider,
-            messages=[{"role": "user", "content": prompt}],
-        )
-        if stop:
-            stop_indexes = (out.find(s) for s in stop if s in out)
-            min_stop = min(stop_indexes, default=-1)
-            if min_stop > -1:
-                out = out[:min_stop]
-        return out
+    def get_messages(self):
+        logger.debug("messages {}", self.messages)
+        return self.messages
diff --git a/pyproject.toml b/pyproject.toml
index 42350b82..8cb3c593 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -25,13 +25,15 @@ build-backend = "poetry.core.masonry.api"
 
 [tool.poetry.dependencies]
 python = "^3.10"
+fastapi = ">=0.95.2"
+uvicorn = ">=0.22.0"
 dynaconf = ">=3.2.0"
 loguru = ">=0.6.0"
 httpx = ">=0.24.1"
 js2py = "^0.74"
-g4f = "0.0.3.0"
-langchain = "0.0.300"
-nest_asyncio = "*"
+g4f = "0.0.3.4"
+# langchain = "0.0.300"
+# nest_asyncio = "*"
 curl-cffi ="0.5.7"
 PyExecJS2="1.6.1"