Xtekky Openai Gpt4 Save

The official gpt4free repository | various collection of powerful language models

Project README

Please provide feedback so this project can be improved, it would be much appreciated
form: https://forms.gle/FeWV9RLEedfdkmFN6

248433934-7886223b-c1d1-4260-82aa-da5741f303bb Written by @xtekky & maintained by @hlohaus

By using this repository or any code related to it, you agree to the legal notice. The author is not responsible for any copies, forks, re-uploads made by other users, or anything else related to GPT4Free. This is the author's only account and repository. To prevent impersonation or irresponsible actions, please comply with the GNU GPL license this Repository uses.

[!Note] Lastet version: PyPI version Docker version
Stats: Downloads Downloads

pip install -U g4f
docker pull hlohaus789/g4f

To do

As per the survey, here is a list of improvements to come

  • Improve Documentation (on g4f.mintlify.app) & Do video tutorials
  • Improve the provider status list & updates
  • Tutorials on how to reverse sites to write your own wrapper (PoC only ofc)
  • Improve the Bing wrapper. (might write a new wrapper in golang as it is very fast)
  • Write a standard provider performance test to improve the stability
  • update the repository to include the new openai library syntax (ex: Openai() class)
  • Potential support and development of local models
  • improve compatibility and error handling

πŸ†• What's New

πŸ“š Table of Contents

πŸ› οΈ Getting Started

Docker container

Quick start:
  1. Download and install Docker
  2. Pull latest image and run the container:
docker pull hlohaus789/g4f
docker run -p 8080:8080 -p 1337:1337 -p 7900:7900 --shm-size="2g" hlohaus789/g4f:latest
  1. Open the included client on: http://localhost:8080/chat/ or set the api base in your client to: http://localhost:1337/v1
  2. (Optional) If you need to log in to a provider, you can view the desktop from the container here: http://localhost:7900/?autoconnect=1&resize=scale&password=secret.

Use python package

  1. Download and install Python (Version 3.10+ is recommended).
  2. Install Google Chrome for providers with webdriver
Install using pypi:

Install all supported tools / all used packages:

pip install -U g4f[all]

Install required packages for the OpenaiChat provider:

pip install -U g4f[openai]

Install required packages for the interference api:

pip install -U g4f[api]

Install required packages for the web interface:

pip install -U g4f[gui]

Install required packages for uploading / generating images:

pip install -U g4f[image]

Install required packages for providers with webdriver:

pip install -U g4f[webdriver]

Install required packages for proxy support:

pip install -U aiohttp_socks
  1. Clone the GitHub repository:
git clone https://github.com/xtekky/gpt4free.git
  1. Navigate to the project directory:
cd gpt4free
  1. (Recommended) Create a Python virtual environment: You can follow the Python official documentation for virtual environments.
python3 -m venv venv
  1. Activate the virtual environment:
    • On Windows:
    • On macOS and Linux:
    source venv/bin/activate
  2. Install minimum requirements:
pip install -r requirements-min.txt
  1. Or install all used Python packages from requirements.txt:
pip install -r requirements.txt
  1. Create a test.py file in the root folder and start using the repo, further Instructions are below
import g4f

Docker for Developers

If you have Docker installed, you can easily set up and run the project without manually installing dependencies.

  1. First, ensure you have both Docker and Docker Compose installed.

  2. Clone the GitHub repo:

git clone https://github.com/xtekky/gpt4free.git
  1. Navigate to the project directory:
cd gpt4free
  1. Build the Docker image:
docker pull selenium/node-chrome
docker-compose build
  1. Start the service using Docker Compose:
docker-compose up

Your server will now be running at http://localhost:1337. You can interact with the API or run your tests as you would normally.

To stop the Docker containers, simply run:

docker-compose down

[!Note] When using Docker, any changes you make to your local files will be reflected in the Docker container thanks to the volume mapping in the docker-compose.yml file. If you add or remove dependencies, however, you'll need to rebuild the Docker image using docker-compose build.

πŸ’‘ Usage

The Web UI

To start the web interface, type the following codes in the command line.

from g4f.gui import run_gui

The g4f Package


import g4f

g4f.debug.logging = True  # Enable debug logging
g4f.debug.version_check = False  # Disable automatic version checking
print(g4f.Provider.Bing.params)  # Print supported args for Bing

# Using automatic a provider for the given model
## Streamed completion
response = g4f.ChatCompletion.create(
    messages=[{"role": "user", "content": "Hello"}],
for message in response:
    print(message, flush=True, end='')

## Normal response
response = g4f.ChatCompletion.create(
    messages=[{"role": "user", "content": "Hello"}],
)  # Alternative model setting

import g4f

allowed_models = [

response = g4f.Completion.create(
    prompt='say this is a test'

import g4f

# Print all available providers
    for provider in g4f.Provider.__providers__
    if provider.working

# Execute with a specific provider
response = g4f.ChatCompletion.create(
    messages=[{"role": "user", "content": "Hello"}],
for message in response:
Cookies / Access Token

For generating images with Bing and for the OpenAi Chat you need cookies or a token from your browser session. From Bing you need the "_U" cookie and from OpenAI you need the "access_token". You can pass the cookies / the access token in the create function or you use the set_cookies setter:

from g4f import set_cookies

set_cookies(".bing", {
  "_U": "cookie value"
set_cookies("chat.openai.com", {
  "access_token": "token value"

from g4f.gui import run_gui

Alternatively, g4f reads the cookies with β€œbrowser_cookie3” from your browser or it starts a browser instance with selenium "webdriver" for logging in. If you use the pip package, you have to install β€œbrowser_cookie3” or "webdriver" by yourself.

pip install browser_cookie3
pip install g4f[webdriver]
Using Browser

Some providers using a browser to bypass the bot protection. They using the selenium webdriver to control the browser. The browser settings and the login data are saved in a custom directory. If the headless mode is enabled, the browser windows are loaded invisibly. For performance reasons, it is recommended to reuse the browser instances and close them yourself at the end:

import g4f
from undetected_chromedriver import Chrome, ChromeOptions
from g4f.Provider import (

options = ChromeOptions()
webdriver = Chrome(options=options, headless=True)
for idx in range(10):
    response = g4f.ChatCompletion.create(
        messages=[{"role": "user", "content": "Suggest me a name."}],
    print(f"{idx}:", response)
Async Support

To enhance speed and overall performance, execute providers asynchronously. The total execution time will be determined by the duration of the slowest provider's execution.

import g4f
import asyncio

_providers = [

async def run_provider(provider: g4f.Provider.BaseProvider):
        response = await g4f.ChatCompletion.create_async(
            messages=[{"role": "user", "content": "Hello"}],
        print(f"{provider.__name__}:", response)
    except Exception as e:
        print(f"{provider.__name__}:", e)
async def run_all():
    calls = [
        run_provider(provider) for provider in _providers
    await asyncio.gather(*calls)

Proxy and Timeout Support

All providers support specifying a proxy and increasing timeout in the create functions.

import g4f

response = g4f.ChatCompletion.create(
    messages=[{"role": "user", "content": "Hello"}],
    # or socks5://user:pass@host:port
    timeout=120,  # in secs

print(f"Result:", response)

You can also set a proxy globally via an environment variable:

export G4F_PROXY="http://host:port"

Interference openai-proxy API (Use with openai python package)

Run interference API from PyPi package

from g4f.api import run_api


Run interference API from repo

If you want to use the embedding function, you need to get a Hugging Face token. You can get one at Hugging Face Tokens. Make sure your role is set to write. If you have your token, just use it instead of the OpenAI api-key.

Run server:

g4f api


python -m g4f.api.run
from openai import OpenAI

client = OpenAI(
    # Set your Hugging Face token as the API key if you use embeddings

    # Set the API base URL if needed, e.g., for a local development environment

def main():
    chat_completion = client.chat.completions.create(
        messages=[{"role": "user", "content": "write a poem about a tree"}],

    if isinstance(chat_completion, dict):
        # Not streaming
        # Streaming
        for token in chat_completion:
            content = token.choices[0].delta.content
            if content is not None:
                print(content, end="", flush=True)

if __name__ == "__main__":

API usage (POST)

Chat completions

Send the POST request to /v1/chat/completions with body containing the model method. This example uses python with requests library:

import requests
url = "http://localhost:1337/v1/chat/completions"
body = {
    "model": "gpt-3.5-turbo-16k",
    "stream": False,
    "messages": [
        {"role": "assistant", "content": "What can you do?"}
json_response = requests.post(url, json=body).json().get('choices', [])

for choice in json_response:
    print(choice.get('message', {}).get('content', ''))

πŸš€ Providers and Models


Website Provider GPT-3.5 GPT-4 Stream Status Auth
bing.com g4f.Provider.Bing ❌ βœ”οΈ βœ”οΈ Active ❌
chat.geekgpt.org g4f.Provider.GeekGpt βœ”οΈ βœ”οΈ βœ”οΈ Inactive ❌
gptchatly.com g4f.Provider.GptChatly βœ”οΈ βœ”οΈ ❌ Unknown ❌
liaobots.site g4f.Provider.Liaobots βœ”οΈ βœ”οΈ βœ”οΈ Unknown ❌
raycast.com g4f.Provider.Raycast βœ”οΈ βœ”οΈ βœ”οΈ Unknown βœ”οΈ


Website Provider GPT-3.5 GPT-4 Stream Status Auth
www.aitianhu.com g4f.Provider.AItianhu βœ”οΈ ❌ βœ”οΈ Unknown ❌
chat3.aiyunos.top g4f.Provider.AItianhuSpace βœ”οΈ ❌ βœ”οΈ Unknown ❌
e.aiask.me g4f.Provider.AiAsk βœ”οΈ ❌ βœ”οΈ Unknown ❌
chat-gpt.org g4f.Provider.Aichat βœ”οΈ ❌ ❌ Unknown ❌
www.chatbase.co g4f.Provider.ChatBase βœ”οΈ ❌ βœ”οΈ Active ❌
chatforai.store g4f.Provider.ChatForAi βœ”οΈ ❌ βœ”οΈ Unknown ❌
chatgpt.ai g4f.Provider.ChatgptAi βœ”οΈ ❌ βœ”οΈ Active ❌
chatgptx.de g4f.Provider.ChatgptX βœ”οΈ ❌ βœ”οΈ Unknown ❌
chat-shared2.zhile.io g4f.Provider.FakeGpt βœ”οΈ ❌ βœ”οΈ Inactive ❌
freegpts1.aifree.site g4f.Provider.FreeGpt βœ”οΈ ❌ βœ”οΈ Active ❌
gptalk.net g4f.Provider.GPTalk βœ”οΈ ❌ βœ”οΈ Active ❌
ai18.gptforlove.com g4f.Provider.GptForLove βœ”οΈ ❌ βœ”οΈ Active ❌
gptgo.ai g4f.Provider.GptGo βœ”οΈ ❌ βœ”οΈ Active ❌
hashnode.com g4f.Provider.Hashnode βœ”οΈ ❌ βœ”οΈ Active ❌
app.myshell.ai g4f.Provider.MyShell βœ”οΈ ❌ βœ”οΈ Unknown ❌
noowai.com g4f.Provider.NoowAi βœ”οΈ ❌ βœ”οΈ Unknown ❌
chat.openai.com g4f.Provider.OpenaiChat βœ”οΈ ❌ βœ”οΈ Unknown βœ”οΈ
theb.ai g4f.Provider.Theb βœ”οΈ ❌ βœ”οΈ Unknown βœ”οΈ
sdk.vercel.ai g4f.Provider.Vercel βœ”οΈ ❌ βœ”οΈ Unknown ❌
you.com g4f.Provider.You βœ”οΈ ❌ βœ”οΈ Active ❌
chat9.yqcloud.top g4f.Provider.Yqcloud βœ”οΈ ❌ βœ”οΈ Unknown ❌
chat.acytoo.com g4f.Provider.Acytoo βœ”οΈ ❌ βœ”οΈ Inactive ❌
aibn.cc g4f.Provider.Aibn βœ”οΈ ❌ βœ”οΈ Inactive ❌
ai.ls g4f.Provider.Ails βœ”οΈ ❌ βœ”οΈ Inactive ❌
chatgpt4online.org g4f.Provider.Chatgpt4Online βœ”οΈ ❌ βœ”οΈ Inactive ❌
chat.chatgptdemo.net g4f.Provider.ChatgptDemo βœ”οΈ ❌ βœ”οΈ Inactive ❌
chatgptduo.com g4f.Provider.ChatgptDuo βœ”οΈ ❌ ❌ Inactive ❌
chatgptfree.ai g4f.Provider.ChatgptFree βœ”οΈ ❌ ❌ Inactive ❌
chatgptlogin.ai g4f.Provider.ChatgptLogin βœ”οΈ ❌ βœ”οΈ Inactive ❌
cromicle.top g4f.Provider.Cromicle βœ”οΈ ❌ βœ”οΈ Inactive ❌
gptgod.site g4f.Provider.GptGod βœ”οΈ ❌ βœ”οΈ Inactive ❌
opchatgpts.net g4f.Provider.Opchatgpts βœ”οΈ ❌ βœ”οΈ Inactive ❌
chat.ylokh.xyz g4f.Provider.Ylokh βœ”οΈ ❌ βœ”οΈ Inactive ❌


Website Provider GPT-3.5 GPT-4 Stream Status Auth
bard.google.com g4f.Provider.Bard ❌ ❌ ❌ Unknown βœ”οΈ
deepinfra.com g4f.Provider.DeepInfra ❌ ❌ βœ”οΈ Active ❌
huggingface.co g4f.Provider.HuggingChat ❌ ❌ βœ”οΈ Active βœ”οΈ
www.llama2.ai g4f.Provider.Llama2 ❌ ❌ βœ”οΈ Unknown ❌
open-assistant.io g4f.Provider.OpenAssistant ❌ ❌ βœ”οΈ Inactive βœ”οΈ


Model Base Provider Provider Website
palm Google g4f.Provider.Bard bard.google.com
h2ogpt-gm-oasst1-en-2048-falcon-7b-v3 Hugging Face g4f.Provider.H2o www.h2o.ai
h2ogpt-gm-oasst1-en-2048-falcon-40b-v1 Hugging Face g4f.Provider.H2o www.h2o.ai
h2ogpt-gm-oasst1-en-2048-open-llama-13b Hugging Face g4f.Provider.H2o www.h2o.ai
claude-instant-v1 Anthropic g4f.Provider.Vercel sdk.vercel.ai
claude-v1 Anthropic g4f.Provider.Vercel sdk.vercel.ai
claude-v2 Anthropic g4f.Provider.Vercel sdk.vercel.ai
command-light-nightly Cohere g4f.Provider.Vercel sdk.vercel.ai
command-nightly Cohere g4f.Provider.Vercel sdk.vercel.ai
gpt-neox-20b Hugging Face g4f.Provider.Vercel sdk.vercel.ai
oasst-sft-1-pythia-12b Hugging Face g4f.Provider.Vercel sdk.vercel.ai
oasst-sft-4-pythia-12b-epoch-3.5 Hugging Face g4f.Provider.Vercel sdk.vercel.ai
santacoder Hugging Face g4f.Provider.Vercel sdk.vercel.ai
bloom Hugging Face g4f.Provider.Vercel sdk.vercel.ai
flan-t5-xxl Hugging Face g4f.Provider.Vercel sdk.vercel.ai
code-davinci-002 OpenAI g4f.Provider.Vercel sdk.vercel.ai
gpt-3.5-turbo-16k OpenAI g4f.Provider.Vercel sdk.vercel.ai
gpt-3.5-turbo-16k-0613 OpenAI g4f.Provider.Vercel sdk.vercel.ai
gpt-4-0613 OpenAI g4f.Provider.Vercel sdk.vercel.ai
text-ada-001 OpenAI g4f.Provider.Vercel sdk.vercel.ai
text-babbage-001 OpenAI g4f.Provider.Vercel sdk.vercel.ai
text-curie-001 OpenAI g4f.Provider.Vercel sdk.vercel.ai
text-davinci-002 OpenAI g4f.Provider.Vercel sdk.vercel.ai
text-davinci-003 OpenAI g4f.Provider.Vercel sdk.vercel.ai
llama13b-v2-chat Replicate g4f.Provider.Vercel sdk.vercel.ai
llama7b-v2-chat Replicate g4f.Provider.Vercel sdk.vercel.ai
🎁 Projects ⭐ Stars πŸ“š Forks πŸ›Ž Issues πŸ“¬ Pull requests
gpt4free Stars Forks Issues Pull Requests
gpt4free-ts Stars Forks Issues Pull Requests
Free AI API's & Potential Providers List Stars Forks Issues Pull Requests
ChatGPT-Clone Stars Forks Issues Pull Requests
ChatGpt Discord Bot Stars Forks Issues Pull Requests
Nyx-Bot (Discord) Stars Forks Issues Pull Requests
LangChain gpt4free Stars Forks Issues Pull Requests
ChatGpt Telegram Bot Stars Forks Issues Pull Requests
ChatGpt Line Bot Stars Forks Issues Pull Requests
Action Translate Readme Stars Forks Issues Pull Requests
Langchain Document GPT Stars Forks Issues Pull Requests

🀝 Contribute

Create Provider with AI Tool

Call in your terminal the create_provider.py script:

python etc/tool/create_provider.py
  1. Enter your name for the new provider.
  2. Copy and paste the cURL command from your browser developer tools.
  3. Let the AI ​​create the provider for you.
  4. Customize the provider according to your needs.

Create Provider

  1. Check out the current list of potential providers, or find your own provider source!
  2. Create a new file in g4f/Provider with the name of the Provider
  3. Implement a class that extends BaseProvider.
from __future__ import annotations

from ..typing import AsyncResult, Messages
from .base_provider import AsyncGeneratorProvider

class HogeService(AsyncGeneratorProvider):
    url                   = "https://chat-gpt.com"
    working               = True
    supports_gpt_35_turbo = True

    async def create_async_generator(
        model: str,
        messages: Messages,
        proxy: str = None,
    ) -> AsyncResult:
        yield ""
  1. Here, you can adjust the settings, for example, if the website does support streaming, set supports_stream to True...
  2. Write code to request the provider in create_async_generator and yield the response, even if it's a one-time response, do not hesitate to look at other providers for inspiration
  3. Add the Provider Name in g4f/Provider/__init__.py
from .HogeService import HogeService

__all__ = [
  1. You are done !, test the provider by calling it:
import g4f

response = g4f.ChatCompletion.create(model='gpt-3.5-turbo', provider=g4f.Provider.PROVIDERNAME,
                                    messages=[{"role": "user", "content": "test"}], stream=g4f.Provider.PROVIDERNAME.supports_stream)

for message in response:
    print(message, flush=True, end='')

πŸ™Œ Contributors

A list of the contributors is available here
The Vercel.py file contains code from vercel-llm-api by @ading2210, which is licensed under the GNU GPL v3
Top 1 Contributor: @hlohaus

This program is licensed under the GNU GPL v3

xtekky/gpt4free: Copyright (C) 2023 xtekky

This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program.  If not, see <https://www.gnu.org/licenses/>.

⭐ Star History

Star History Chart

πŸ“„ License

This project is licensed under GNU_GPL_v3.0.

(πŸ”Ό Back to top)

Open Source Agenda is not affiliated with "Xtekky Openai Gpt4" Project. README Source: xtekky/gpt4free

Open Source Agenda Badge

Open Source Agenda Rating