19 files changed, 673 insertions, 308 deletions

diff --git a/g4f/Provider/Bestim.py b/g4f/Provider/Bestim.py
index 312655b8..be95b48a 100644
--- a/g4f/Provider/Bestim.py
+++ b/g4f/Provider/Bestim.py
@@ -1,42 +1,35 @@
 from __future__         import annotations

 

-from ..typing           import Messages, List, Dict

+from ..typing           import Messages

 from .base_provider     import BaseProvider, CreateResult

+from ..requests         import get_session_from_browser

 from uuid               import uuid4

 import requests

 

-def format_prompt(messages) -> List[Dict[str, str]]:

-

-    return [{"id": str(uuid4()), "content": '\n'.join(f'{m["role"]}: {m["content"]}' for m in messages), "from": "you"}]

-

 class Bestim(BaseProvider):

     url = "https://chatgpt.bestim.org"

     supports_gpt_35_turbo = True

     supports_message_history = True

-    working = True

+    working = False

     supports_stream = True

 

-    @staticmethod

+    @classmethod

     def create_completion(

+        cls,

         model: str,

         messages: Messages,

         stream: bool,

         proxy: str = None, 

         **kwargs

     ) -> CreateResult:

-        

+        session = get_session_from_browser(cls.url, proxy=proxy)

         headers = {

-            'POST': '/chat/send2/ HTTP/3',

-            'Host': 'chatgpt.bestim.org',

             'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:121.0) Gecko/20100101 Firefox/121.0',

             'Accept': 'application/json, text/event-stream',

             'Accept-Language': 'en-US,en;q=0.5',

             'Accept-Encoding': 'gzip, deflate, br',

             'Referer': 'https://chatgpt.bestim.org/chat/',

-            'Content-Type': 'application/json',

-            'Content-Length': '109',

             'Origin': 'https://chatgpt.bestim.org',

-            'Cookie': 'NpZAER=qKkRHguMIOraVbJAWpoyzGLFjZwYlm; qKkRHguMIOraVbJAWpoyzGLFjZwYlm=8ebb5ae1561bde05354de5979b52c6e1-1704058188-1704058188; NpZAER_hits=2; _csrf-front=fcf20965823c0a152ae8f9cdf15b23022bb26cdc6bf32a9d4c8bfe78dcc6b807a%3A2%3A%7Bi%3A0%3Bs%3A11%3A%22_csrf-front%22%3Bi%3A1%3Bs%3A32%3A%22a5wP6azsc7dxV8rmwAXaNsl8XS1yvW5V%22%3B%7D',

             'Alt-Used': 'chatgpt.bestim.org',

             'Connection': 'keep-alive',

             'Sec-Fetch-Dest': 'empty',

@@ -44,27 +37,25 @@ class Bestim(BaseProvider):
             'Sec-Fetch-Site': 'same-origin',

             'TE': 'trailers'

         }

-

         data = {

-

-            "messagesHistory": format_prompt(messages),

+            "messagesHistory": [{

+                "id": str(uuid4()),

+                "content": m["content"],

+                "from": "you" if m["role"] == "user" else "bot"

+            } for m in messages],

             "type": "chat",

         }

-            

-        response = requests.post(

+        response = session.post(

             url="https://chatgpt.bestim.org/chat/send2/",

             headers=headers,

             json=data,

-            proxies={"https": proxy}

+            proxies={"https": proxy},

+            stream=True

         )

-

         response.raise_for_status()

-

-        for chunk in response.iter_lines():

-

-            if b"event: trylimit" not in chunk:

-

-                yield chunk.decode().removeprefix("data: ")

+        for line in response.iter_lines():

+            if not line.startswith(b"event: trylimit"):

+                yield line.decode().removeprefix("data: ")

 

 

 

diff --git a/g4f/Provider/Bing.py b/g4f/Provider/Bing.py
index e03b413f..b2e10e5d 100644
--- a/g4f/Provider/Bing.py
+++ b/g4f/Provider/Bing.py
@@ -1,35 +1,25 @@
 from __future__ import annotations
 
-import string
 import random
 import json
 import os
-import re
-import io
-import base64
-import numpy as np
 import uuid
-import urllib.parse
 import time
-from PIL import Image
-from aiohttp        import ClientSession, ClientTimeout
-from ..typing       import AsyncResult, Messages
+from urllib import parse
+from aiohttp import ClientSession, ClientTimeout
+
+from ..typing import AsyncResult, Messages
 from .base_provider import AsyncGeneratorProvider
+from ..webdriver import get_browser, get_driver_cookies
+from .bing.upload_image import upload_image
+from .bing.create_images import create_images, format_images_markdown, wait_for_login
+from .bing.conversation import Conversation, create_conversation, delete_conversation
 
 class Tones():
     creative = "Creative"
     balanced = "Balanced"
     precise = "Precise"
 
-default_cookies = {
-    'SRCHD'         : 'AF=NOFORM',
-    'PPLState'      : '1',
-    'KievRPSSecAuth': '',
-    'SUID'          : '',
-    'SRCHUSR'       : '',
-    'SRCHHPGUSR'    : f'HV={int(time.time())}',
-}
-
 class Bing(AsyncGeneratorProvider):
     url = "https://bing.com/chat"
     working = True
@@ -55,9 +45,9 @@ class Bing(AsyncGeneratorProvider):
             context = create_context(messages[:-1])
         
         if not cookies:
-            cookies = default_cookies
+            cookies = Defaults.cookies
         else:
-            for key, value in default_cookies.items():
+            for key, value in Defaults.cookies.items():
                 if key not in cookies:
                     cookies[key] = value
 
@@ -71,106 +61,6 @@ def create_context(messages: Messages):
         for message in messages
     )
 
-class Conversation():
-    def __init__(self, conversationId: str, clientId: str, conversationSignature: str, imageInfo: dict=None) -> None:
-        self.conversationId = conversationId
-        self.clientId = clientId
-        self.conversationSignature = conversationSignature
-        self.imageInfo = imageInfo
-
-async def create_conversation(session: ClientSession, tone: str, image: str = None, proxy: str = None) -> Conversation:
-    url = 'https://www.bing.com/turing/conversation/create?bundleVersion=1.1199.4'
-    async with session.get(url, proxy=proxy) as response:
-        data = await response.json()
-
-        conversationId = data.get('conversationId')
-        clientId = data.get('clientId')
-        conversationSignature = response.headers.get('X-Sydney-Encryptedconversationsignature')
-
-        if not conversationId or not clientId or not conversationSignature:
-            raise Exception('Failed to create conversation.')
-        conversation = Conversation(conversationId, clientId, conversationSignature, None)
-        if isinstance(image,str):
-            try:
-                config = {
-                    "visualSearch": {
-                        "maxImagePixels": 360000,
-                        "imageCompressionRate": 0.7,
-                        "enableFaceBlurDebug": 0,
-                    }
-                }
-                is_data_uri_an_image(image)
-                img_binary_data = extract_data_uri(image)
-                is_accepted_format(img_binary_data)
-                img = Image.open(io.BytesIO(img_binary_data))
-                width, height = img.size
-                max_image_pixels = config['visualSearch']['maxImagePixels']
-                compression_rate = config['visualSearch']['imageCompressionRate']
-
-                if max_image_pixels / (width * height) < 1:
-                    new_width = int(width * np.sqrt(max_image_pixels / (width * height)))
-                    new_height = int(height * np.sqrt(max_image_pixels / (width * height)))
-                else:
-                    new_width = width
-                    new_height = height
-                try:
-                    orientation = get_orientation(img)
-                except Exception:
-                    orientation = None
-                new_img = process_image(orientation, img, new_width, new_height)
-                new_img_binary_data = compress_image_to_base64(new_img, compression_rate)
-                data, boundary = build_image_upload_api_payload(new_img_binary_data, conversation, tone)
-                headers = session.headers.copy()
-                headers["content-type"] = f'multipart/form-data; boundary={boundary}'
-                headers["referer"] = 'https://www.bing.com/search?q=Bing+AI&showconv=1&FORM=hpcodx'
-                headers["origin"] = 'https://www.bing.com'
-                async with session.post("https://www.bing.com/images/kblob", data=data, headers=headers, proxy=proxy) as image_upload_response:
-                    if image_upload_response.status != 200:
-                        raise Exception("Failed to upload image.")
-
-                    image_info = await image_upload_response.json()
-                    if not image_info.get('blobId'):
-                        raise Exception("Failed to parse image info.")
-                    result = {'bcid': image_info.get('blobId', "")}
-                    result['blurredBcid'] = image_info.get('processedBlobId', "")
-                    if result['blurredBcid'] != "":
-                        result["imageUrl"] = "https://www.bing.com/images/blob?bcid=" + result['blurredBcid']
-                    elif result['bcid'] != "":
-                        result["imageUrl"] = "https://www.bing.com/images/blob?bcid=" + result['bcid']
-                    result['originalImageUrl'] = (
-                        "https://www.bing.com/images/blob?bcid="
-                        + result['blurredBcid']
-                        if config['visualSearch']["enableFaceBlurDebug"]
-                        else "https://www.bing.com/images/blob?bcid="
-                        + result['bcid']
-                    )
-                    conversation.imageInfo = result
-            except Exception as e:
-                print(f"An error happened while trying to send image: {str(e)}")
-        return conversation
-
-async def list_conversations(session: ClientSession) -> list:
-    url = "https://www.bing.com/turing/conversation/chats"
-    async with session.get(url) as response:
-        response = await response.json()
-        return response["chats"]
-        
-async def delete_conversation(session: ClientSession, conversation: Conversation, proxy: str = None) -> list:
-    url = "https://sydney.bing.com/sydney/DeleteSingleConversation"
-    json = {
-        "conversationId": conversation.conversationId,
-        "conversationSignature": conversation.conversationSignature,
-        "participant": {"id": conversation.clientId},
-        "source": "cib",
-        "optionsSets": ["autosave"]
-    }
-    async with session.post(url, json=json, proxy=proxy) as response:
-        try:
-            response = await response.json()
-            return response["result"]["value"] == "Success"
-        except:
-            return False
-
 class Defaults:
     delimiter = "\x1e"
     ip_address = f"13.{random.randint(104, 107)}.{random.randint(0, 255)}.{random.randint(0, 255)}"
@@ -264,123 +154,28 @@ class Defaults:
         'eredirecturl',
         'nojbfedge'
     ]
+    
+    cookies = {
+        'SRCHD'         : 'AF=NOFORM',
+        'PPLState'      : '1',
+        'KievRPSSecAuth': '',
+        'SUID'          : '',
+        'SRCHUSR'       : '',
+        'SRCHHPGUSR'    : f'HV={int(time.time())}',
+    }
 
 def format_message(msg: dict) -> str:
     return json.dumps(msg, ensure_ascii=False) + Defaults.delimiter
 
-def build_image_upload_api_payload(image_bin: str, conversation: Conversation, tone: str):
-    payload = {
-        'invokedSkills': ["ImageById"],
-        'subscriptionId': "Bing.Chat.Multimodal",
-        'invokedSkillsRequestData': {
-            'enableFaceBlur': True
-        },
-        'convoData': {
-            'convoid': "",
-            'convotone': tone
-        }
-    }
-    knowledge_request = {
-        'imageInfo': {},
-        'knowledgeRequest': payload
-    }
-    boundary="----WebKitFormBoundary" + ''.join(random.choices(string.ascii_letters + string.digits, k=16))
-    data = (
-        f'--{boundary}'
-        + '\r\nContent-Disposition: form-data; name="knowledgeRequest"\r\n\r\n'
-        + json.dumps(knowledge_request, ensure_ascii=False)
-        + "\r\n--"
-        + boundary
-        + '\r\nContent-Disposition: form-data; name="imageBase64"\r\n\r\n'
-        + image_bin
-        + "\r\n--"
-        + boundary
-        + "--\r\n"
-    )
-    return data, boundary
-
-def is_data_uri_an_image(data_uri: str):
-    try:
-        # Check if the data URI starts with 'data:image' and contains an image format (e.g., jpeg, png, gif)
-        if not re.match(r'data:image/(\w+);base64,', data_uri):
-            raise ValueError("Invalid data URI image.")
-            # Extract the image format from the data URI
-        image_format = re.match(r'data:image/(\w+);base64,', data_uri).group(1)
-        # Check if the image format is one of the allowed formats (jpg, jpeg, png, gif)
-        if image_format.lower() not in ['jpeg', 'jpg', 'png', 'gif']:
-            raise ValueError("Invalid image format (from mime file type).")
-    except Exception as e:
-        raise e
-
-def is_accepted_format(binary_data: bytes) -> bool:
-        try:
-            check = False
-            if binary_data.startswith(b'\xFF\xD8\xFF'):
-                check = True  # It's a JPEG image
-            elif binary_data.startswith(b'\x89PNG\r\n\x1a\n'):
-                check = True  # It's a PNG image
-            elif binary_data.startswith(b'GIF87a') or binary_data.startswith(b'GIF89a'):
-                check = True  # It's a GIF image
-            elif binary_data.startswith(b'\x89JFIF') or binary_data.startswith(b'JFIF\x00'):
-                check = True  # It's a JPEG image
-            elif binary_data.startswith(b'\xFF\xD8'):
-                check = True  # It's a JPEG image
-            elif binary_data.startswith(b'RIFF') and binary_data[8:12] == b'WEBP':
-                check = True  # It's a WebP image
-            # else we raise ValueError
-            if not check:
-                raise ValueError("Invalid image format (from magic code).")
-        except Exception as e:
-            raise e
-    
-def extract_data_uri(data_uri: str) -> bytes:
-    try:
-        data = data_uri.split(",")[1]
-        data = base64.b64decode(data)
-        return data
-    except Exception as e:
-        raise e
-
-def get_orientation(data: bytes) -> int:
-    try:
-        if data[:2] != b'\xFF\xD8':
-            raise Exception('NotJpeg')
-        with Image.open(data) as img:
-            exif_data = img._getexif()
-            if exif_data is not None:
-                orientation = exif_data.get(274)  # 274 corresponds to the orientation tag in EXIF
-                if orientation is not None:
-                    return orientation
-    except Exception:
-        pass
-
-def process_image(orientation: int, img: Image.Image, new_width: int, new_height: int) -> Image.Image:
-    try:
-        # Initialize the canvas
-        new_img = Image.new("RGB", (new_width, new_height), color="#FFFFFF")
-        if orientation:
-            if orientation > 4:
-                img = img.transpose(Image.FLIP_LEFT_RIGHT)
-            if orientation in [3, 4]:
-                img = img.transpose(Image.ROTATE_180)
-            if orientation in [5, 6]:
-                img = img.transpose(Image.ROTATE_270)
-            if orientation in [7, 8]:
-                img = img.transpose(Image.ROTATE_90)
-        new_img.paste(img, (0, 0))
-        return new_img
-    except Exception as e:
-        raise e
-    
-def compress_image_to_base64(img, compression_rate) -> str:
-    try:
-        output_buffer = io.BytesIO()
-        img.save(output_buffer, format="JPEG", quality=int(compression_rate * 100))
-        return base64.b64encode(output_buffer.getvalue()).decode('utf-8')
-    except Exception as e:
-        raise e
-
-def create_message(conversation: Conversation, prompt: str, tone: str, context: str = None, web_search: bool = False, gpt4_turbo: bool = False) -> str:
+def create_message(
+    conversation: Conversation,
+    prompt: str,
+    tone: str,
+    context: str = None,
+    image_info: dict = None,
+    web_search: bool = False,
+    gpt4_turbo: bool = False
+) -> str:
     options_sets = Defaults.optionsSets
     if tone == Tones.creative:
         options_sets.append("h3imaginative")
@@ -429,9 +224,9 @@ def create_message(conversation: Conversation, prompt: str, tone: str, context:
         'target': 'chat',
         'type': 4
     }
-    if conversation.imageInfo != None and "imageUrl" in conversation.imageInfo and "originalImageUrl" in conversation.imageInfo:
-        struct['arguments'][0]['message']['originalImageUrl'] = conversation.imageInfo['originalImageUrl']
-        struct['arguments'][0]['message']['imageUrl'] = conversation.imageInfo['imageUrl']
+    if image_info and "imageUrl" in image_info and "originalImageUrl" in image_info:
+        struct['arguments'][0]['message']['originalImageUrl'] = image_info['originalImageUrl']
+        struct['arguments'][0]['message']['imageUrl'] = image_info['imageUrl']
         struct['arguments'][0]['experienceType'] = None
         struct['arguments'][0]['attachedFileInfo'] = {"fileName": None, "fileType": None}
     if context:
@@ -454,28 +249,39 @@ async def stream_generate(
         web_search: bool = False,
         gpt4_turbo: bool = False
     ):
+    headers = Defaults.headers
+    if cookies:
+        headers["Cookie"] = "; ".join(f"{k}={v}" for k, v in cookies.items())
     async with ClientSession(
-            timeout=ClientTimeout(total=900),
-            headers=Defaults.headers if not cookies else {**Defaults.headers, "Cookie": "; ".join(f"{k}={v}" for k, v in cookies.items())},
-        ) as session:
-        conversation = await create_conversation(session, tone, image, proxy)
+        timeout=ClientTimeout(total=900),
+        headers=headers
+    ) as session:
+        conversation = await create_conversation(session, proxy)
+        image_info = None
+        if image:
+            image_info = await upload_image(session, image, tone, proxy)
         try:
-            async with session.ws_connect('wss://sydney.bing.com/sydney/ChatHub', autoping=False, params={'sec_access_token': conversation.conversationSignature}, proxy=proxy) as wss:
-
+            async with session.ws_connect(
+                'wss://sydney.bing.com/sydney/ChatHub',
+                autoping=False,
+                params={'sec_access_token': conversation.conversationSignature},
+                proxy=proxy
+            ) as wss:
                 await wss.send_str(format_message({'protocol': 'json', 'version': 1}))
                 await wss.receive(timeout=900)
-                await wss.send_str(create_message(conversation, prompt, tone, context, web_search, gpt4_turbo))
+                await wss.send_str(create_message(conversation, prompt, tone, context, image_info, web_search, gpt4_turbo))
 
                 response_txt = ''
                 returned_text = ''
                 final = False
                 while not final:
                     msg = await wss.receive(timeout=900)
+                    if not msg.data:
+                        continue
                     objects = msg.data.split(Defaults.delimiter)
                     for obj in objects:
                         if obj is None or not obj:
                             continue
-
                         response = json.loads(obj)
                         if response.get('type') == 1 and response['arguments'][0].get('messages'):
                             message = response['arguments'][0]['messages'][0]
@@ -488,9 +294,11 @@ async def stream_generate(
                                         inline_txt = card['inlines'][0].get('text')
                                         response_txt += inline_txt + '\n'
                                 elif message.get('contentType') == "IMAGE":
-                                    query = urllib.parse.quote(message.get('text'))
-                                    url = f"\nhttps://www.bing.com/images/create?q={query}"
-                                    response_txt += url
+                                    prompt = message.get('text')
+                                    try:
+                                        response_txt += format_images_markdown(await create_images(session, prompt, proxy), prompt)
+                                    except:
+                                        response_txt += f"\nhttps://www.bing.com/images/create?q={parse.quote(prompt)}"
                                     final = True
                             if response_txt.startswith(returned_text):
                                 new = response_txt[len(returned_text):]
@@ -500,7 +308,18 @@ async def stream_generate(
                         elif response.get('type') == 2:
                             result = response['item']['result']
                             if result.get('error'):
-                                raise Exception(f"{result['value']}: {result['message']}")
+                                if result["value"] == "CaptchaChallenge":
+                                    driver = get_browser(proxy=proxy)
+                                    try:
+                                        for chunk in wait_for_login(driver):
+                                            yield chunk
+                                        cookies = get_driver_cookies(driver)
+                                    finally:
+                                        driver.quit()
+                                    async for chunk in stream_generate(prompt, tone, image, context, proxy, cookies, web_search, gpt4_turbo):
+                                        yield chunk
+                                else:
+                                    raise Exception(f"{result['value']}: {result['message']}")
                             return
         finally:
             await delete_conversation(session, conversation, proxy)
diff --git a/g4f/Provider/FreeChatgpt.py b/g4f/Provider/FreeChatgpt.py
new file mode 100644
index 00000000..05a0016e
--- /dev/null
+++ b/g4f/Provider/FreeChatgpt.py
@@ -0,0 +1,68 @@
+from __future__ import annotations
+
+import json
+from aiohttp import ClientSession
+
+from ..typing import AsyncResult, Messages
+from .base_provider import AsyncGeneratorProvider
+
+
+class FreeChatgpt(AsyncGeneratorProvider):
+    url = "https://free.chatgpt.org.uk"
+    working = True
+    supports_gpt_35_turbo = True
+    supports_gpt_4 = True
+    supports_message_history = True
+
+    @classmethod
+    async def create_async_generator(
+        cls,
+        model: str,
+        messages: Messages,
+        proxy: str = None,
+        **kwargs
+    ) -> AsyncResult:
+        if not model:
+            model = "gpt-3.5-turbo"
+        headers = {
+            "User-Agent": "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/119.0",
+            "Accept": "application/json, text/event-stream",
+            "Accept-Language": "de,en-US;q=0.7,en;q=0.3",
+            "Accept-Encoding": "gzip, deflate, br",
+            "Content-Type": "application/json",
+            "Referer": "https://free.chatgpt.org.uk/",
+            "x-requested-with": "XMLHttpRequest",
+            "Origin": "https://free.chatgpt.org.uk",
+            "Sec-Fetch-Dest": "empty",
+            "Sec-Fetch-Mode": "cors",
+            "Sec-Fetch-Site": "same-origin",
+            "Connection": "keep-alive",
+            "Alt-Used": "free.chatgpt.org.uk",
+            "Pragma": "no-cache",
+            "Cache-Control": "no-cache",
+            "TE": "trailers",
+        }
+        async with ClientSession(headers=headers) as session:
+            data = {
+                "messages": messages,
+                "isAzure": False,
+                "azureApiVersion": "2023-08-01-preview",
+                "stream": True,
+                "model": model,
+                "temperature": 0.5,
+                "presence_penalty": 0,
+                "frequency_penalty": 0,
+                "top_p": 1,
+                "baseUrl": "/api/openai",
+                "maxIterations": 10,
+                "returnIntermediateSteps": True,
+                "useTools": ["web-search", "calculator", "web-browser"],
+                **kwargs
+            }
+            async with session.post(f"{cls.url}/api/langchain/tool/agent/nodejs", json=data, proxy=proxy) as response:
+                response.raise_for_status()
+                async for line in response.content:
+                    if line.startswith(b"data: "):
+                        data = json.loads(line[6:])
+                        if data["isSuccess"] and not data["isToolMessage"]:
+                            yield data["message"]
\ No newline at end of file
diff --git a/g4f/Provider/__init__.py b/g4f/Provider/__init__.py
index 19212836..7a3dbb76 100644
--- a/g4f/Provider/__init__.py
+++ b/g4f/Provider/__init__.py
@@ -3,6 +3,7 @@ from __future__ import annotations
 from ..base_provider  import BaseProvider, ProviderType
 from .retry_provider  import RetryProvider
 from .base_provider   import AsyncProvider, AsyncGeneratorProvider
+from .create_images   import CreateImagesProvider
 from .deprecated      import *
 from .needs_auth      import *
 from .unfinished      import *
@@ -30,6 +31,7 @@ from .ChatgptX        import ChatgptX
 from .Chatxyz         import Chatxyz
 from .DeepInfra       import DeepInfra
 from .FakeGpt         import FakeGpt
+from .FreeChatgpt     import FreeChatgpt
 from .FreeGpt         import FreeGpt
 from .GeekGpt         import GeekGpt
 from .GeminiProChat   import GeminiProChat
diff --git a/g4f/Provider/bing/conversation.py b/g4f/Provider/bing/conversation.py
new file mode 100644
index 00000000..ef45cd82
--- /dev/null
+++ b/g4f/Provider/bing/conversation.py
@@ -0,0 +1,43 @@
+from aiohttp import ClientSession
+
+
+class Conversation():
+    def __init__(self, conversationId: str, clientId: str, conversationSignature: str) -> None:
+        self.conversationId = conversationId
+        self.clientId = clientId
+        self.conversationSignature = conversationSignature
+
+async def create_conversation(session: ClientSession, proxy: str = None) -> Conversation:
+    url = 'https://www.bing.com/turing/conversation/create?bundleVersion=1.1199.4'
+    async with session.get(url, proxy=proxy) as response:
+        data = await response.json()
+
+        conversationId = data.get('conversationId')
+        clientId = data.get('clientId')
+        conversationSignature = response.headers.get('X-Sydney-Encryptedconversationsignature')
+
+        if not conversationId or not clientId or not conversationSignature:
+            raise Exception('Failed to create conversation.')
+        return Conversation(conversationId, clientId, conversationSignature)
+        
+async def list_conversations(session: ClientSession) -> list:
+    url = "https://www.bing.com/turing/conversation/chats"
+    async with session.get(url) as response:
+        response = await response.json()
+        return response["chats"]
+        
+async def delete_conversation(session: ClientSession, conversation: Conversation, proxy: str = None) -> list:
+    url = "https://sydney.bing.com/sydney/DeleteSingleConversation"
+    json = {
+        "conversationId": conversation.conversationId,
+        "conversationSignature": conversation.conversationSignature,
+        "participant": {"id": conversation.clientId},
+        "source": "cib",
+        "optionsSets": ["autosave"]
+    }
+    try:
+        async with session.post(url, json=json, proxy=proxy) as response:
+            response = await response.json()
+            return response["result"]["value"] == "Success"
+    except:
+        return False
\ No newline at end of file
diff --git a/g4f/Provider/bing/create_images.py b/g4f/Provider/bing/create_images.py
new file mode 100644
index 00000000..b203a0dc
--- /dev/null
+++ b/g4f/Provider/bing/create_images.py
@@ -0,0 +1,164 @@
+import asyncio
+import time, json, os
+from aiohttp import ClientSession
+from bs4 import BeautifulSoup
+from urllib.parse import quote
+from typing import Generator
+
+from ..create_images import CreateImagesProvider
+from ..helper import get_cookies, get_event_loop
+from ...webdriver import WebDriver, get_driver_cookies, get_browser
+from ...base_provider import ProviderType
+
+BING_URL = "https://www.bing.com"
+
+def wait_for_login(driver: WebDriver, timeout: int = 1200) -> None:
+    driver.get(f"{BING_URL}/")
+    value = driver.get_cookie("_U")
+    if value:
+        return
+    start_time = time.time()
+    while True:
+        if time.time() - start_time > timeout:
+            raise RuntimeError("Timeout error")
+        value = driver.get_cookie("_U")
+        if value:
+            return
+        time.sleep(0.5)
+
+def create_session(cookies: dict) -> ClientSession:
+    headers = {
+        "accept": "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.7",
+        "accept-encoding": "gzip, deflate, br",
+        "accept-language": "en-US,en;q=0.9,zh-CN;q=0.8,zh-TW;q=0.7,zh;q=0.6",
+        "content-type": "application/x-www-form-urlencoded",
+        "referrer-policy": "origin-when-cross-origin",
+        "referrer": "https://www.bing.com/images/create/",
+        "origin": "https://www.bing.com",
+        "user-agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/111.0.0.0 Safari/537.36 Edg/111.0.1661.54",
+        "sec-ch-ua": "\"Microsoft Edge\";v=\"111\", \"Not(A:Brand\";v=\"8\", \"Chromium\";v=\"111\"",
+        "sec-ch-ua-mobile": "?0",
+        "sec-fetch-dest": "document",
+        "sec-fetch-mode": "navigate",
+        "sec-fetch-site": "same-origin",
+        "sec-fetch-user": "?1",
+        "upgrade-insecure-requests": "1",
+    }
+    if cookies:
+        headers["cookie"] = "; ".join(f"{k}={v}" for k, v in cookies.items())
+    return ClientSession(headers=headers)
+
+async def create_images(session: ClientSession, prompt: str, proxy: str = None, timeout: int = 300) -> list:
+    url_encoded_prompt = quote(prompt)    
+    payload = f"q={url_encoded_prompt}&rt=4&FORM=GENCRE"
+    url = f"{BING_URL}/images/create?q={url_encoded_prompt}&rt=4&FORM=GENCRE"
+    async with session.post(
+        url,
+        allow_redirects=False,
+        data=payload,
+        timeout=timeout,
+    ) as response:
+        response.raise_for_status()
+        errors = [
+            "this prompt is being reviewed",
+            "this prompt has been blocked",
+            "we're working hard to offer image creator in more languages"
+        ]
+        text = (await response.text()).lower()
+        for error in errors:
+            if error in text:
+                raise RuntimeError(f"Create images failed: {error}")
+    if response.status != 302:
+        url = f"{BING_URL}/images/create?q={url_encoded_prompt}&rt=3&FORM=GENCRE"
+        async with session.post(url, allow_redirects=False, proxy=proxy, timeout=timeout) as response:
+            if response.status != 302:
+                raise RuntimeError(f"Create images failed. Status Code: {response.status}")
+
+    redirect_url = response.headers["Location"].replace("&nfy=1", "")
+    redirect_url = f"{BING_URL}{redirect_url}"
+    request_id = redirect_url.split("id=")[1]
+    async with session.get(redirect_url) as response:
+        response.raise_for_status()
+
+    polling_url = f"{BING_URL}/images/create/async/results/{request_id}?q={url_encoded_prompt}"
+    start_time = time.time()
+    while True:
+        if time.time() - start_time > timeout:
+            raise RuntimeError(f"Timeout error after {timeout} seconds")
+        async with session.get(polling_url) as response:
+            if response.status != 200:
+                raise RuntimeError(f"Polling images faild. Status Code: {response.status}")
+            text = await response.text()
+            if not text:
+                await asyncio.sleep(1)
+            else:
+                break
+    error = None
+    try:
+        error = json.loads(text).get("errorMessage")
+    except:
+        pass
+    if error == "Pending":
+        raise RuntimeError("Prompt is been blocked")
+    elif error:
+        raise RuntimeError(error)
+    return read_images(text)
+
+def read_images(text: str) -> list:
+    html_soup = BeautifulSoup(text, "html.parser")
+    tags = html_soup.find_all("img")
+    image_links = [img["src"] for img in tags if "mimg" in img["class"]]
+    images = [link.split("?w=")[0] for link in image_links]
+    bad_images = [
+        "https://r.bing.com/rp/in-2zU3AJUdkgFe7ZKv19yPBHVs.png",
+        "https://r.bing.com/rp/TX9QuO3WzcCJz1uaaSwQAz39Kb0.jpg",
+    ]
+    if any(im in bad_images for im in images):
+        raise RuntimeError("Bad images found")
+    if not images:
+        raise RuntimeError("No images found")
+    return images
+
+def format_images_markdown(images: list, prompt: str) -> str:
+    images = [f"[![#{idx+1} {prompt}]({image}?w=200&h=200)]({image})" for idx, image in enumerate(images)]
+    images = "\n".join(images)
+    start_flag = "<!-- generated images start -->\n"
+    end_flag = "<!-- generated images end -->\n"
+    return f"\n{start_flag}{images}\n{end_flag}\n"
+
+async def create_images_markdown(cookies: dict, prompt: str, proxy: str = None) -> str:
+    session = create_session(cookies)
+    try:
+        images = await create_images(session, prompt, proxy)
+        return format_images_markdown(images, prompt)
+    finally:
+        await session.close()
+
+def get_cookies_from_browser(proxy: str = None) -> dict:
+    driver = get_browser(proxy=proxy)
+    try:
+        wait_for_login(driver)
+        return get_driver_cookies(driver)
+    finally:
+        driver.quit()
+
+def create_completion(prompt: str, cookies: dict = None, proxy: str = None) -> Generator:
+    loop = get_event_loop()
+    if not cookies:
+        cookies = get_cookies(".bing.com")
+    if "_U" not in cookies:
+        login_url = os.environ.get("G4F_LOGIN_URL")
+        if login_url:
+            yield f"Please login: [Bing]({login_url})\n\n"
+        cookies = get_cookies_from_browser(proxy)
+    yield loop.run_until_complete(create_images_markdown(cookies, prompt, proxy))
+
+async def create_async(prompt: str, cookies: dict = None, proxy: str = None) -> str:
+    if not cookies:
+        cookies = get_cookies(".bing.com")
+    if "_U" not in cookies:
+        cookies = get_cookies_from_browser(proxy)
+    return await create_images_markdown(cookies, prompt, proxy)
+
+def patch_provider(provider: ProviderType) -> CreateImagesProvider:
+    return CreateImagesProvider(provider, create_completion, create_async)
\ No newline at end of file
diff --git a/g4f/Provider/bing/upload_image.py b/g4f/Provider/bing/upload_image.py
new file mode 100644
index 00000000..329e6df4
--- /dev/null
+++ b/g4f/Provider/bing/upload_image.py
@@ -0,0 +1,162 @@
+from __future__ import annotations
+
+import string
+import random
+import json
+import re
+import io
+import base64
+import numpy as np
+from PIL import Image
+from aiohttp import ClientSession
+
+async def upload_image(
+    session: ClientSession,
+    image: str,
+    tone: str,
+    proxy: str = None
+):
+    try:
+        image_config = {
+            "maxImagePixels": 360000,
+            "imageCompressionRate": 0.7,
+            "enableFaceBlurDebug": 0,
+        }
+        is_data_uri_an_image(image)
+        img_binary_data = extract_data_uri(image)
+        is_accepted_format(img_binary_data)
+        img = Image.open(io.BytesIO(img_binary_data))
+        width, height = img.size
+        max_image_pixels = image_config['maxImagePixels']
+        if max_image_pixels / (width * height) < 1:
+            new_width = int(width * np.sqrt(max_image_pixels / (width * height)))
+            new_height = int(height * np.sqrt(max_image_pixels / (width * height)))
+        else:
+            new_width = width
+            new_height = height
+        try:
+            orientation = get_orientation(img)
+        except Exception:
+            orientation = None
+        new_img = process_image(orientation, img, new_width, new_height)
+        new_img_binary_data = compress_image_to_base64(new_img, image_config['imageCompressionRate'])
+        data, boundary = build_image_upload_api_payload(new_img_binary_data, tone)
+        headers = session.headers.copy()
+        headers["content-type"] = f'multipart/form-data; boundary={boundary}'
+        headers["referer"] = 'https://www.bing.com/search?q=Bing+AI&showconv=1&FORM=hpcodx'
+        headers["origin"] = 'https://www.bing.com'
+        async with session.post("https://www.bing.com/images/kblob", data=data, headers=headers, proxy=proxy) as response:
+            if response.status != 200:
+                raise RuntimeError("Failed to upload image.")
+            image_info = await response.json()
+            if not image_info.get('blobId'):
+                raise RuntimeError("Failed to parse image info.")
+            result = {'bcid': image_info.get('blobId', "")}
+            result['blurredBcid'] = image_info.get('processedBlobId', "")
+            if result['blurredBcid'] != "":
+                result["imageUrl"] = "https://www.bing.com/images/blob?bcid=" + result['blurredBcid']
+            elif result['bcid'] != "":
+                result["imageUrl"] = "https://www.bing.com/images/blob?bcid=" + result['bcid']
+            result['originalImageUrl'] = (
+                "https://www.bing.com/images/blob?bcid="
+                + result['blurredBcid']
+                if image_config["enableFaceBlurDebug"]
+                else "https://www.bing.com/images/blob?bcid="
+                + result['bcid']
+            )
+            return result
+    except Exception as e:
+        raise RuntimeError(f"Upload image failed: {e}")
+    
+
+def build_image_upload_api_payload(image_bin: str, tone: str):
+    payload = {
+        'invokedSkills': ["ImageById"],
+        'subscriptionId': "Bing.Chat.Multimodal",
+        'invokedSkillsRequestData': {
+            'enableFaceBlur': True
+        },
+        'convoData': {
+            'convoid': "",
+            'convotone': tone
+        }
+    }
+    knowledge_request = {
+        'imageInfo': {},
+        'knowledgeRequest': payload
+    }
+    boundary="----WebKitFormBoundary" + ''.join(random.choices(string.ascii_letters + string.digits, k=16))
+    data = (
+        f'--{boundary}'
+        + '\r\nContent-Disposition: form-data; name="knowledgeRequest"\r\n\r\n'
+        + json.dumps(knowledge_request, ensure_ascii=False)
+        + "\r\n--"
+        + boundary
+        + '\r\nContent-Disposition: form-data; name="imageBase64"\r\n\r\n'
+        + image_bin
+        + "\r\n--"
+        + boundary
+        + "--\r\n"
+    )
+    return data, boundary
+
+def is_data_uri_an_image(data_uri: str):
+    # Check if the data URI starts with 'data:image' and contains an image format (e.g., jpeg, png, gif)
+    if not re.match(r'data:image/(\w+);base64,', data_uri):
+        raise ValueError("Invalid data URI image.")
+        # Extract the image format from the data URI
+    image_format = re.match(r'data:image/(\w+);base64,', data_uri).group(1)
+    # Check if the image format is one of the allowed formats (jpg, jpeg, png, gif)
+    if image_format.lower() not in ['jpeg', 'jpg', 'png', 'gif']:
+        raise ValueError("Invalid image format (from mime file type).")
+
+def is_accepted_format(binary_data: bytes) -> bool:
+    if binary_data.startswith(b'\xFF\xD8\xFF'):
+        pass # It's a JPEG image
+    elif binary_data.startswith(b'\x89PNG\r\n\x1a\n'):
+        pass # It's a PNG image
+    elif binary_data.startswith(b'GIF87a') or binary_data.startswith(b'GIF89a'):
+        pass # It's a GIF image
+    elif binary_data.startswith(b'\x89JFIF') or binary_data.startswith(b'JFIF\x00'):
+        pass # It's a JPEG image
+    elif binary_data.startswith(b'\xFF\xD8'):
+        pass # It's a JPEG image
+    elif binary_data.startswith(b'RIFF') and binary_data[8:12] == b'WEBP':
+        pass # It's a WebP image
+    else:
+        raise ValueError("Invalid image format (from magic code).")
+    
+def extract_data_uri(data_uri: str) -> bytes:
+    data = data_uri.split(",")[1]
+    data = base64.b64decode(data)
+    return data
+
+def get_orientation(data: bytes) -> int:
+    if data[:2] != b'\xFF\xD8':
+        raise Exception('NotJpeg')
+    with Image.open(data) as img:
+        exif_data = img._getexif()
+        if exif_data is not None:
+            orientation = exif_data.get(274)  # 274 corresponds to the orientation tag in EXIF
+            if orientation is not None:
+                return orientation
+
+def process_image(orientation: int, img: Image.Image, new_width: int, new_height: int) -> Image.Image:
+    # Initialize the canvas
+    new_img = Image.new("RGB", (new_width, new_height), color="#FFFFFF")
+    if orientation:
+        if orientation > 4:
+            img = img.transpose(Image.FLIP_LEFT_RIGHT)
+        if orientation in [3, 4]:
+            img = img.transpose(Image.ROTATE_180)
+        if orientation in [5, 6]:
+            img = img.transpose(Image.ROTATE_270)
+        if orientation in [7, 8]:
+            img = img.transpose(Image.ROTATE_90)
+    new_img.paste(img, (0, 0))
+    return new_img
+    
+def compress_image_to_base64(image: Image.Image, compression_rate: float) -> str:
+    output_buffer = io.BytesIO()
+    image.save(output_buffer, format="JPEG", quality=int(compression_rate * 100))
+    return base64.b64encode(output_buffer.getvalue()).decode('utf-8')
\ No newline at end of file
diff --git a/g4f/Provider/create_images.py b/g4f/Provider/create_images.py
new file mode 100644
index 00000000..9c76a742
--- /dev/null
+++ b/g4f/Provider/create_images.py
@@ -0,0 +1,84 @@
+from __future__ import annotations
+
+import re
+import asyncio
+from ..typing import CreateResult, Messages
+from ..base_provider import BaseProvider, ProviderType
+
+system_message = """
+You can generate custom images with the DALL-E 3 image generator.
+To generate a image with a prompt, do this:
+<img data-prompt=\"keywords for the image\">
+Don't use images with data uri. It is important to use a prompt instead.
+<img data-prompt=\"image caption\">
+"""
+
+class CreateImagesProvider(BaseProvider):
+    def __init__(
+        self,
+        provider: ProviderType,
+        create_images: callable,
+        create_async: callable,
+        system_message: str = system_message,
+        include_placeholder: bool = True
+    ) -> None:
+        self.provider = provider
+        self.create_images = create_images
+        self.create_images_async = create_async
+        self.system_message = system_message
+        self.__name__ = provider.__name__
+        self.working = provider.working
+        self.supports_stream = provider.supports_stream
+        self.include_placeholder = include_placeholder
+        if hasattr(provider, "url"):
+            self.url = provider.url
+
+    def create_completion(
+        self,
+        model: str,
+        messages: Messages,
+        stream: bool = False,
+        **kwargs
+    ) -> CreateResult:
+        messages.insert(0, {"role": "system", "content": self.system_message})
+        buffer = ""
+        for chunk in self.provider.create_completion(model, messages, stream, **kwargs):
+            if buffer or "<" in chunk:
+                buffer += chunk
+                if ">" in buffer:
+                    match = re.search(r'<img data-prompt="(.*?)">', buffer)
+                    if match:
+                        placeholder, prompt = match.group(0), match.group(1)
+                        start, append = buffer.split(placeholder, 1)
+                        if start:
+                            yield start
+                        if self.include_placeholder:
+                            yield placeholder
+                        yield from self.create_images(prompt)
+                        if append:
+                            yield append
+                    else:
+                        yield buffer
+                    buffer = ""
+            else:
+                yield chunk
+
+    async def create_async(
+        self,
+        model: str,
+        messages: Messages,
+        **kwargs
+    ) -> str:
+        messages.insert(0, {"role": "system", "content": self.system_message})
+        response = await self.provider.create_async(model, messages, **kwargs)
+        matches = re.findall(r'(<img data-prompt="(.*?)">)', result)
+        results = []
+        for _, prompt in matches:
+            results.append(self.create_images_async(prompt))
+        results = await asyncio.gather(*results)
+        for idx, result in enumerate(results):
+            placeholder = matches[idx][0]
+            if self.include_placeholder:
+                result = placeholder + result
+            response = response.replace(placeholder, result)
+        return result
\ No newline at end of file
diff --git a/g4f/Provider/needs_auth/Bard.py b/g4f/Provider/needs_auth/Bard.py
index cf1000c4..aea67874 100644
--- a/g4f/Provider/needs_auth/Bard.py
+++ b/g4f/Provider/needs_auth/Bard.py
@@ -64,11 +64,14 @@ XMLHttpRequest.prototype.open = function(method, url) {
 """
             driver.execute_script(script)
 
-            # Submit prompt
-            driver.find_element(By.CSS_SELECTOR, "div.ql-editor.textarea").send_keys(prompt)
-            driver.find_element(By.CSS_SELECTOR, "div.ql-editor.textarea").send_keys(Keys.ENTER)
-
-            # Yield response
+            textarea = driver.find_element(By.CSS_SELECTOR, "div.ql-editor.textarea")
+            lines = prompt.splitlines()
+            for idx, line in enumerate(lines):
+                textarea.send_keys(line)
+                if (len(lines) - 1 != idx):
+                    textarea.send_keys(Keys.SHIFT + "\n")
+            textarea.send_keys(Keys.ENTER)
+            
             while True:
                 chunk = driver.execute_script("return window._message;")
                 if chunk:
diff --git a/g4f/__init__.py b/g4f/__init__.py
index 699dc238..dc7808f9 100644
--- a/g4f/__init__.py
+++ b/g4f/__init__.py
@@ -68,6 +68,7 @@ class ChatCompletion:
                ignored  : list[str] = None, 
                ignore_working: bool = False,
                ignore_stream_and_auth: bool = False,
+               patch_provider: callable = None,
                **kwargs) -> Union[CreateResult, str]:
 
         model, provider = get_model_and_provider(model, provider, stream, ignored, ignore_working, ignore_stream_and_auth)
@@ -83,6 +84,9 @@ class ChatCompletion:
             if proxy:
                 kwargs['proxy'] = proxy
 
+        if patch_provider:
+            provider = patch_provider(provider)
+
         result = provider.create_completion(model, messages, stream, **kwargs)
         return result if stream else ''.join(result)
 
@@ -92,6 +96,7 @@ class ChatCompletion:
                      provider : Union[ProviderType, str, None] = None,
                      stream   : bool = False,
                      ignored  : list[str] = None,
+                     patch_provider: callable = None,
                      **kwargs) -> Union[AsyncResult, str]:
 
         model, provider = get_model_and_provider(model, provider, False, ignored)
@@ -101,6 +106,9 @@ class ChatCompletion:
                 return provider.create_async_generator(model, messages, **kwargs)
             raise StreamNotSupportedError(f'{provider.__name__} does not support "stream" argument in "create_async"')
 
+        if patch_provider:
+            provider = patch_provider(provider)
+
         return provider.create_async(model, messages, **kwargs)
 
 class Completion:
diff --git a/g4f/gui/client/html/index.html b/g4f/gui/client/html/index.html
index da7aeefb..b47f2a37 100644
--- a/g4f/gui/client/html/index.html
+++ b/g4f/gui/client/html/index.html
@@ -125,6 +125,11 @@
                     <span class="about">Web Access</span>
                 </div>
                 <div class="field">
+                    <input type="checkbox" id="patch" />
+                    <label for="patch" title="Works only with Bing and some other providers"></label>
+                    <span class="about">Image Generator</span>
+                </div>
+                <div class="field">
                     <select name="model" id="model">
                         <option value="gpt-3.5-turbo" selected="">gpt-3.5-turbo</option>
                         <option value="gpt-3.5-turbo-0613">gpt-3.5-turbo-0613</option>
diff --git a/g4f/gui/client/js/chat.v1.js b/g4f/gui/client/js/chat.v1.js
index a335a3cc..9d49d24e 100644
--- a/g4f/gui/client/js/chat.v1.js
+++ b/g4f/gui/client/js/chat.v1.js
@@ -20,7 +20,10 @@ message_input.addEventListener("focus", () => {
 });
 
 const markdown_render = (content) => {
-    return markdown.render(content)
+    return markdown.render(content
+        .replaceAll(/<!--.+-->/gm, "")
+        .replaceAll(/<img data-prompt="[^>]+">/gm, "")
+    )
         .replaceAll("<a href=", '<a target="_blank" href=')
         .replaceAll('<code>', '<code class="language-plaintext">')
 }
@@ -68,6 +71,15 @@ const ask_gpt = async () => {
     regenerate.classList.add(`regenerate-hidden`);
     messages = await get_messages(window.conversation_id);
 
+    // Remove generated images from history
+    for (i in messages) {
+        messages[i]["content"] = messages[i]["content"].replace(
+            /<!-- generated images start -->[\s\S]+<!-- generated images end -->/m,
+            ""
+        )
+        delete messages[i]["provider"];
+    }
+
     window.scrollTo(0, 0);
     window.controller = new AbortController();
 
@@ -91,7 +103,8 @@ const ask_gpt = async () => {
             </div>
             <div class="content" id="gpt_${window.token}">
                 <div class="provider"></div>
-                <div class="content_inner"><div id="cursor"></div></div>
+                <div class="content_inner"></div>
+                <div id="cursor"></div>
             </div>
         </div>
     `;
@@ -115,6 +128,7 @@ const ask_gpt = async () => {
                 jailbreak: jailbreak.options[jailbreak.selectedIndex].value,
                 internet_access: document.getElementById(`switch`).checked,
                 provider: provider.options[provider.selectedIndex].value,
+                patch_provider: document.getElementById('patch').checked,
                 meta: {
                     id: window.token,
                     content: {
@@ -163,8 +177,6 @@ const ask_gpt = async () => {
     } catch (e) {
         console.log(e);
 
-        let cursorDiv = document.getElementById(`cursor`);
-        if (cursorDiv) cursorDiv.parentNode.removeChild(cursorDiv);
 
         if (e.name != `AbortError`) {
             text = `oops ! something went wrong, please try again / reload. [stacktrace in console]`;
@@ -174,6 +186,8 @@ const ask_gpt = async () => {
             text += ` [aborted]`
         }
     }
+    let cursorDiv = document.getElementById(`cursor`);
+    if (cursorDiv) cursorDiv.parentNode.removeChild(cursorDiv);
     add_message(window.conversation_id, "assistant", text, provider);
     message_box.scrollTop = message_box.scrollHeight;
     await remove_cancel_button();
@@ -430,7 +444,7 @@ document.querySelector(".mobile-sidebar").addEventListener("click", (event) => {
 });
 
 const register_settings_localstorage = async () => {
-    settings_ids = ["switch", "model", "jailbreak"];
+    settings_ids = ["switch", "model", "jailbreak", "patch", "provider"];
     settings_elements = settings_ids.map((id) => document.getElementById(id));
     settings_elements.map((element) =>
         element.addEventListener(`change`, async (event) => {
@@ -449,7 +463,7 @@ const register_settings_localstorage = async () => {
 };
 
 const load_settings_localstorage = async () => {
-    settings_ids = ["switch", "model", "jailbreak"];
+    settings_ids = ["switch", "model", "jailbreak", "patch", "provider"];
     settings_elements = settings_ids.map((id) => document.getElementById(id));
     settings_elements.map((element) => {
         if (localStorage.getItem(element.id)) {
diff --git a/g4f/gui/server/backend.py b/g4f/gui/server/backend.py
index 595f5aa1..67f13de4 100644
--- a/g4f/gui/server/backend.py
+++ b/g4f/gui/server/backend.py
@@ -1,11 +1,11 @@
 import logging
-import g4f
-from g4f.Provider import __providers__
-
 import json
-from flask      import request, Flask
-from .internet  import get_search_message
-from g4f import debug, version
+from flask import request, Flask
+from g4f import debug, version, models
+from g4f import _all_models, get_last_provider, ChatCompletion
+from g4f.Provider import __providers__
+from g4f.Provider.bing.create_images import patch_provider
+from .internet import get_search_message
 
 debug.logging = True
 
@@ -45,7 +45,7 @@ class Backend_Api:
         return 'ok', 200
     
     def models(self):
-        return g4f._all_models
+        return _all_models
     
     def providers(self):
         return [
@@ -69,25 +69,27 @@ class Backend_Api:
         if request.json.get('internet_access'):
             messages[-1]["content"] = get_search_message(messages[-1]["content"])
         model = request.json.get('model')
-        model = model if model else g4f.models.default
+        model = model if model else models.default
         provider = request.json.get('provider', '').replace('g4f.Provider.', '')
         provider = provider if provider and provider != "Auto" else None
-            
+        patch = patch_provider if request.json.get('patch_provider') else None
+
         def try_response():
             try:
                 first = True
-                for chunk in g4f.ChatCompletion.create(
+                for chunk in ChatCompletion.create(
                     model=model,
                     provider=provider,
                     messages=messages,
                     stream=True,
-                    ignore_stream_and_auth=True
+                    ignore_stream_and_auth=True,
+                    patch_provider=patch
                 ):
                     if first:
                         first = False
                         yield json.dumps({
                             'type'    : 'provider',
-                            'provider': g4f.get_last_provider(True)
+                            'provider': get_last_provider(True)
                         }) + "\n"
                     yield json.dumps({
                         'type'   : 'content',
diff --git a/g4f/gui/server/internet.py b/g4f/gui/server/internet.py
index 97f842d7..6c2e3a89 100644
--- a/g4f/gui/server/internet.py
+++ b/g4f/gui/server/internet.py
@@ -145,5 +145,5 @@ User request:
 """
         return message
     except Exception as e:
-        print("Couldn't search DuckDuckGo:", e)
+        print("Couldn't do web search:", e)
         return prompt
diff --git a/g4f/models.py b/g4f/models.py
index 6103ca03..21e85c72 100644
--- a/g4f/models.py
+++ b/g4f/models.py
@@ -7,20 +7,18 @@ from .Provider   import (
     ChatgptNext,
     HuggingChat,
     ChatgptDemo,
+    FreeChatgpt,
     GptForLove,
     ChatgptAi,
     DeepInfra,
-    OnlineGpt,
     ChatBase,
     Liaobots,
     GeekGpt,
     FakeGpt,
     FreeGpt,
-    Berlin,
     Llama2,
     Vercel, 
     Phind,
-    Koala,
     GptGo,
     Gpt6,
     Bard, 
@@ -58,13 +56,12 @@ gpt_35_long = Model(
     best_provider = RetryProvider([
         FreeGpt, You,
         GeekGpt, FakeGpt,
-        Berlin, Koala,
         Chatgpt4Online,
         ChatgptDemoAi,
-        OnlineGpt,
         ChatgptNext,
         ChatgptDemo,
         Gpt6,
+        FreeChatgpt,
     ])
 )
 
diff --git a/g4f/requests.py b/g4f/requests.py
index 00ab9488..467ea371 100644
--- a/g4f/requests.py
+++ b/g4f/requests.py
@@ -6,7 +6,7 @@ from functools import partialmethod
 from typing import AsyncGenerator
 from urllib.parse import urlparse
 from curl_cffi.requests import AsyncSession, Session, Response
-from .webdriver import WebDriver, WebDriverSession, bypass_cloudflare
+from .webdriver import WebDriver, WebDriverSession, bypass_cloudflare, get_driver_cookies
 
 class StreamResponse:
     def __init__(self, inner: Response) -> None:
@@ -56,8 +56,7 @@ class StreamSession(AsyncSession):
 def get_session_from_browser(url: str, webdriver: WebDriver = None, proxy: str = None, timeout: int = 120):
     with WebDriverSession(webdriver, "", proxy=proxy, virtual_display=True) as driver:
         bypass_cloudflare(driver, url, timeout)
-
-        cookies = dict([(cookie["name"], cookie["value"]) for cookie in driver.get_cookies()])
+        cookies = get_driver_cookies(driver)
         user_agent = driver.execute_script("return navigator.userAgent")
 
     parse = urlparse(url)
diff --git a/g4f/version.py b/g4f/version.py
index 9e572bbb..44d14369 100644
--- a/g4f/version.py
+++ b/g4f/version.py
@@ -40,7 +40,7 @@ class VersionUtils():
     def check_pypi_version(self) -> None:
         try:
             if self.current_version != self.latest_version:
-                print(f'New pypi version: {self.latest_version} (current: {self.version}) | pip install -U g4f')
+                print(f'New pypi version: {self.latest_version} (current: {self.current_version}) | pip install -U g4f')
         except Exception as e:
             print(f'Failed to check g4f pypi version: {e}')
          
diff --git a/g4f/webdriver.py b/g4f/webdriver.py
index d274c619..da283409 100644
--- a/g4f/webdriver.py
+++ b/g4f/webdriver.py
@@ -39,6 +39,9 @@ def get_browser(
         headless=headless
     )
 
+def get_driver_cookies(driver: WebDriver):
+    return dict([(cookie["name"], cookie["value"]) for cookie in driver.get_cookies()])
+
 def bypass_cloudflare(driver: WebDriver, url: str, timeout: int) -> None:
     # Open website
     driver.get(url)
diff --git a/requirements.txt b/requirements.txt
index 27067e7a..b9212b4e 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -26,4 +26,5 @@ async-property
 undetected-chromedriver
 asyncstdlib
 async_property
-beautifulsoup4
+brotli
+beautifulsoup4
\ No newline at end of file