> By using this repository or any code related to it, you agree to the [legal notice](LEGAL_NOTICE.md). 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] > Latest pypi version: [`0.1.8.7`](https://pypi.org/project/g4f/0.1.8.7) ```sh pip install -U g4f ``` ## 🆕 What's New - Join our Telegram Channel: [t.me/g4f_channel](https://telegram.me/g4f_channel) - Join our Discord Group: [discord.gg/XfybzPXPH5](https://discord.gg/XfybzPXPH5) - Explore the g4f Documentation (unfinished): [g4f.mintlify.app](https://g4f.mintlify.app) | Contribute to the docs via: [github.com/xtekky/gpt4free-docs](https://github.com/xtekky/gpt4free-docs) ## 📚 Table of Contents - [🆕 What's New](#-whats-new) - [📚 Table of Contents](#-table-of-contents) - [🛠️ Getting Started](#️-getting-started) - [Prerequisites:](#prerequisites) - [Setting up the project:](#setting-up-the-project) - [Install using pypi](#install-using-pypi) - [or](#or) - [Setting up with Docker:](#setting-up-with-docker) - [💡 Usage](#-usage) - [The `g4f` Package](#the-g4f-package) - [ChatCompletion](#chatcompletion) - [Completion](#completion) - [Providers](#providers) - [Cookies Required](#cookies-required) - [Async Support](#async-support) - [Proxy and Timeout Support](#proxy-and-timeout-support) - [Interference openai-proxy API (Use with openai python package)](#interference-openai-proxy-api-use-with-openai-python-package) - [Run interference API from PyPi package](#run-interference-api-from-pypi-package) - [Run interference API from repo](#run-interference-api-from-repo) - [🚀 Providers and Models](#-providers-and-models) - [GPT-4](#gpt-4) - [GPT-3.5](#gpt-35) - [Other](#other) - [Models](#models) - [🔗 Related GPT4Free Projects](#-related-gpt4free-projects) - [🤝 Contribute](#-contribute) - [Create Provider with AI Tool](#create-provider-with-ai-tool) - [Create Provider](#create-provider) - [🙌 Contributors](#-contributors) - [©️ Copyright](#️-copyright) - [⭐ Star History](#-star-history) - [📄 License](#-license) ## 🛠️ Getting Started #### Prerequisites: 1. [Download and install Python](https://www.python.org/downloads/) (Version 3.10+ is recommended). #### Setting up the project: ##### Install using pypi ``` pip install -U g4f ``` ##### or 1. Clone the GitHub repository: ``` git clone https://github.com/xtekky/gpt4free.git ``` 2. Navigate to the project directory: ``` cd gpt4free ``` 3. (Recommended) Create a Python virtual environment: You can follow the [Python official documentation](https://docs.python.org/3/tutorial/venv.html) for virtual environments. ``` python3 -m venv venv ``` 4. Activate the virtual environment: - On Windows: ``` .\venv\Scripts\activate ``` - On macOS and Linux: ``` source venv/bin/activate ``` 5. Install the required Python packages from `requirements.txt`: ``` pip install -r requirements.txt ``` 6. Create a `test.py` file in the root folder and start using the repo, further Instructions are below ```py import g4f ... ``` ##### Setting up with Docker: 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. - [Install Docker](https://docs.docker.com/get-docker/) - [Install Docker Compose](https://docs.docker.com/compose/install/) 2. Clone the GitHub repo: ```bash git clone https://github.com/xtekky/gpt4free.git ``` 3. Navigate to the project directory: ```bash cd gpt4free ``` 4. Build the Docker image: ```bash docker-compose build ``` 5. Start the service using Docker Compose: ```bash 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: ```bash 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`. ## 💡 Usage ### The `g4f` Package #### ChatCompletion ```python import g4f g4f.debug.logging = True # Enable logging g4f.check_version = False # Disable automatic version checking print(g4f.version) # Check version print(g4f.Provider.Ails.params) # Supported args # Automatic selection of provider # Streamed completion response = g4f.ChatCompletion.create( model="gpt-3.5-turbo", messages=[{"role": "user", "content": "Hello"}], stream=True, ) for message in response: print(message, flush=True, end='') # Normal response response = g4f.ChatCompletion.create( model=g4f.models.gpt_4, messages=[{"role": "user", "content": "Hello"}], ) # Alternative model setting print(response) ``` ##### Completion ```python import g4f allowed_models = [ 'code-davinci-002', 'text-ada-001', 'text-babbage-001', 'text-curie-001', 'text-davinci-002', 'text-davinci-003' ] response = g4f.Completion.create( model='text-davinci-003', prompt='say this is a test' ) print(response) ``` ##### Providers ```python import g4f from g4f.Provider import ( AItianhu, Aichat, Bard, Bing, ChatBase, ChatgptAi, OpenaiChat, Vercel, You, Yqcloud, ) # Set with provider response = g4f.ChatCompletion.create( model="gpt-3.5-turbo", provider=g4f.Provider.Aichat, messages=[{"role": "user", "content": "Hello"}], stream=True, ) for message in response: print(message) ``` ##### Using Browser Some providers using a 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: ```python import g4f from undetected_chromedriver import Chrome, ChromeOptions from g4f.Provider import ( Bard, Poe, AItianhuSpace, MyShell, Phind, PerplexityAi, ) options = ChromeOptions() options.add_argument("--incognito"); webdriver = Chrome(options=options, headless=True) for idx in range(10): response = g4f.ChatCompletion.create( model=g4f.models.default, provider=g4f.Provider.Phind, messages=[{"role": "user", "content": "Suggest me a name."}], webdriver=webdriver ) print(f"{idx}:", response) webdriver.quit() ``` ##### Cookies Required Cookies are essential for the proper functioning of some service providers. It is imperative to maintain an active session, typically achieved by logging into your account. When running the g4f package locally, the package automatically retrieves cookies from your web browser using the `get_cookies` function. However, if you're not running it locally, you'll need to provide the cookies manually by passing them as parameters using the `cookies` parameter. ```python import g4f from g4f.Provider import ( Bing, HuggingChat, OpenAssistant, ) # Usage response = g4f.ChatCompletion.create( model=g4f.models.default, messages=[{"role": "user", "content": "Hello"}], provider=Bing, #cookies=g4f.get_cookies(".google.com"), cookies={"cookie_name": "value", "cookie_name2": "value2"}, auth=True ) ``` ##### 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. ```python import g4f import asyncio _providers = [ g4f.Provider.Aichat, g4f.Provider.ChatBase, g4f.Provider.Bing, g4f.Provider.GptGo, g4f.Provider.You, g4f.Provider.Yqcloud, ] async def run_provider(provider: g4f.Provider.BaseProvider): try: response = await g4f.ChatCompletion.create_async( model=g4f.models.default, messages=[{"role": "user", "content": "Hello"}], provider=provider, ) 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) asyncio.run(run_all()) ``` ##### Proxy and Timeout Support All providers support specifying a proxy and increasing timeout in the create functions. ```python import g4f response = g4f.ChatCompletion.create( model=g4f.models.default, messages=[{"role": "user", "content": "Hello"}], proxy="http://host:port", # or socks5://user:pass@host:port timeout=120, # in secs ) print(f"Result:", response) ``` ### Interference openai-proxy API (Use with openai python package) #### Run interference API from PyPi package ```python from g4f.api import run_api 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](https://huggingface.co/settings/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: ```sh g4f api ``` or ```sh python -m g4f.api.run ``` ```python import openai # Set your Hugging Face token as the API key if you use embeddings # If you don't use embeddings, leave it empty openai.api_key = "YOUR_HUGGING_FACE_TOKEN" # Replace with your actual token # Set the API base URL if needed, e.g., for a local development environment openai.api_base = "http://localhost:1337/v1" def main(): chat_completion = openai.ChatCompletion.create( model="gpt-3.5-turbo", messages=[{"role": "user", "content": "write a poem about a tree"}], stream=True, ) if isinstance(chat_completion, dict): # Not streaming print(chat_completion.choices[0].message.content) else: # Streaming for token in chat_completion: content = token["choices"][0]["delta"].get("content") if content is not None: print(content, end="", flush=True) if __name__ == "__main__": main() ``` ## 🚀 Providers and Models ### GPT-4 | Website | Provider | GPT-3.5 | GPT-4 | Stream | Status | Auth | | ------ | ------- | ------- | ----- | ------ | ------ | ---- | | [bing.com](https://bing.com/chat) | `g4f.Provider.Bing` | ❌ | ✔️ | ✔️ |  | ❌ | | [chat.geekgpt.org](https://chat.geekgpt.org) | `g4f.Provider.GeekGpt` | ✔️ | ✔️ | ✔️ |  | ❌ | | [gptchatly.com](https://gptchatly.com) | `g4f.Provider.GptChatly` | ✔️ | ✔️ | ❌ |  | ❌ | | [liaobots.site](https://liaobots.site) | `g4f.Provider.Liaobots` | ✔️ | ✔️ | ✔️ |  | ❌ | | [www.phind.com](https://www.phind.com) | `g4f.Provider.Phind` | ❌ | ✔️ | ✔️ |  | ❌ | | [raycast.com](https://raycast.com) | `g4f.Provider.Raycast` | ✔️ | ✔️ | ✔️ |  | ✔️ | ### GPT-3.5 | Website | Provider | GPT-3.5 | GPT-4 | Stream | Status | Auth | | ------ | ------- | ------- | ----- | ------ | ------ | ---- | | [www.aitianhu.com](https://www.aitianhu.com) | `g4f.Provider.AItianhu` | ✔️ | ❌ | ✔️ |  | ❌ | | [chat3.aiyunos.top](https://chat3.aiyunos.top/) | `g4f.Provider.AItianhuSpace` | ✔️ | ❌ | ✔️ |  | ❌ | | [e.aiask.me](https://e.aiask.me) | `g4f.Provider.AiAsk` | ✔️ | ❌ | ✔️ |  | ❌ | | [chat-gpt.org](https://chat-gpt.org/chat) | `g4f.Provider.Aichat` | ✔️ | ❌ | ❌ |  | ❌ | | [www.chatbase.co](https://www.chatbase.co) | `g4f.Provider.ChatBase` | ✔️ | ❌ | ✔️ |  | ❌ | | [chatforai.store](https://chatforai.store) | `g4f.Provider.ChatForAi` | ✔️ | ❌ | ✔️ |  | ❌ | | [chatgpt.ai](https://chatgpt.ai) | `g4f.Provider.ChatgptAi` | ✔️ | ❌ | ✔️ |  | ❌ | | [chatgptx.de](https://chatgptx.de) | `g4f.Provider.ChatgptX` | ✔️ | ❌ | ✔️ |  | ❌ | | [chat-shared2.zhile.io](https://chat-shared2.zhile.io) | `g4f.Provider.FakeGpt` | ✔️ | ❌ | ✔️ |  | ❌ | | [freegpts1.aifree.site](https://freegpts1.aifree.site/) | `g4f.Provider.FreeGpt` | ✔️ | ❌ | ✔️ |  | ❌ | | [gptalk.net](https://gptalk.net) | `g4f.Provider.GPTalk` | ✔️ | ❌ | ✔️ |  | ❌ | | [ai18.gptforlove.com](https://ai18.gptforlove.com) | `g4f.Provider.GptForLove` | ✔️ | ❌ | ✔️ |  | ❌ | | [gptgo.ai](https://gptgo.ai) | `g4f.Provider.GptGo` | ✔️ | ❌ | ✔️ |  | ❌ | | [hashnode.com](https://hashnode.com) | `g4f.Provider.Hashnode` | ✔️ | ❌ | ✔️ |  | ❌ | | [app.myshell.ai](https://app.myshell.ai/chat) | `g4f.Provider.MyShell` | ✔️ | ❌ | ✔️ |  | ❌ | | [noowai.com](https://noowai.com) | `g4f.Provider.NoowAi` | ✔️ | ❌ | ✔️ |  | ❌ | | [chat.openai.com](https://chat.openai.com) | `g4f.Provider.OpenaiChat` | ✔️ | ❌ | ✔️ |  | ✔️ | | [theb.ai](https://theb.ai) | `g4f.Provider.Theb` | ✔️ | ❌ | ✔️ |  | ✔️ | | [sdk.vercel.ai](https://sdk.vercel.ai) | `g4f.Provider.Vercel` | ✔️ | ❌ | ✔️ |  | ❌ | | [you.com](https://you.com) | `g4f.Provider.You` | ✔️ | ❌ | ✔️ |  | ❌ | | [chat9.yqcloud.top](https://chat9.yqcloud.top/) | `g4f.Provider.Yqcloud` | ✔️ | ❌ | ✔️ |  | ❌ | | [chat.acytoo.com](https://chat.acytoo.com) | `g4f.Provider.Acytoo` | ✔️ | ❌ | ✔️ |  | ❌ | | [aibn.cc](https://aibn.cc) | `g4f.Provider.Aibn` | ✔️ | ❌ | ✔️ |  | ❌ | | [ai.ls](https://ai.ls) | `g4f.Provider.Ails` | ✔️ | ❌ | ✔️ |  | ❌ | | [chatgpt4online.org](https://chatgpt4online.org) | `g4f.Provider.Chatgpt4Online` | ✔️ | ❌ | ✔️ |  | ❌ | | [chat.chatgptdemo.net](https://chat.chatgptdemo.net) | `g4f.Provider.ChatgptDemo` | ✔️ | ❌ | ✔️ |  | ❌ | | [chatgptduo.com](https://chatgptduo.com) | `g4f.Provider.ChatgptDuo` | ✔️ | ❌ | ❌ |  | ❌ | | [chatgptfree.ai](https://chatgptfree.ai) | `g4f.Provider.ChatgptFree` | ✔️ | ❌ | ❌ |  | ❌ | | [chatgptlogin.ai](https://chatgptlogin.ai) | `g4f.Provider.ChatgptLogin` | ✔️ | ❌ | ✔️ |  | ❌ | | [cromicle.top](https://cromicle.top) | `g4f.Provider.Cromicle` | ✔️ | ❌ | ✔️ |  | ❌ | | [gptgod.site](https://gptgod.site) | `g4f.Provider.GptGod` | ✔️ | ❌ | ✔️ |  | ❌ | | [opchatgpts.net](https://opchatgpts.net) | `g4f.Provider.Opchatgpts` | ✔️ | ❌ | ✔️ |  | ❌ | | [chat.ylokh.xyz](https://chat.ylokh.xyz) | `g4f.Provider.Ylokh` | ✔️ | ❌ | ✔️ |  | ❌ | ### Other | Website | Provider | GPT-3.5 | GPT-4 | Stream | Status | Auth | | ------ | ------- | ------- | ----- | ------ | ------ | ---- | | [bard.google.com](https://bard.google.com) | `g4f.Provider.Bard` | ❌ | ❌ | ❌ |  | ✔️ | | [deepinfra.com](https://deepinfra.com) | `g4f.Provider.DeepInfra` | ❌ | ❌ | ✔️ |  | ❌ | | [huggingface.co](https://huggingface.co/chat) | `g4f.Provider.HuggingChat` | ❌ | ❌ | ✔️ |  | ✔️ | | [www.llama2.ai](https://www.llama2.ai) | `g4f.Provider.Llama2` | ❌ | ❌ | ✔️ |  | ❌ | | [open-assistant.io](https://open-assistant.io/chat) | `g4f.Provider.OpenAssistant` | ❌ | ❌ | ✔️ |  | ✔️ | ### Models | Model | Base Provider | Provider | Website | | --------------------------------------- | ------------- | ------------------- | ------------------------------------------- | | palm | Google | g4f.Provider.Bard | [bard.google.com](https://bard.google.com/) | | h2ogpt-gm-oasst1-en-2048-falcon-7b-v3 | Hugging Face | g4f.Provider.H2o | [www.h2o.ai](https://www.h2o.ai/) | | h2ogpt-gm-oasst1-en-2048-falcon-40b-v1 | Hugging Face | g4f.Provider.H2o | [www.h2o.ai](https://www.h2o.ai/) | | h2ogpt-gm-oasst1-en-2048-open-llama-13b | Hugging Face | g4f.Provider.H2o | [www.h2o.ai](https://www.h2o.ai/) | | claude-instant-v1 | Anthropic | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | claude-v1 | Anthropic | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | claude-v2 | Anthropic | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | command-light-nightly | Cohere | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | command-nightly | Cohere | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | gpt-neox-20b | Hugging Face | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | oasst-sft-1-pythia-12b | Hugging Face | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | oasst-sft-4-pythia-12b-epoch-3.5 | Hugging Face | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | santacoder | Hugging Face | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | bloom | Hugging Face | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | flan-t5-xxl | Hugging Face | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | code-davinci-002 | OpenAI | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | gpt-3.5-turbo-16k | OpenAI | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | gpt-3.5-turbo-16k-0613 | OpenAI | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | gpt-4-0613 | OpenAI | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | text-ada-001 | OpenAI | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | text-babbage-001 | OpenAI | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | text-curie-001 | OpenAI | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | text-davinci-002 | OpenAI | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | text-davinci-003 | OpenAI | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | llama13b-v2-chat | Replicate | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | | llama7b-v2-chat | Replicate | g4f.Provider.Vercel | [sdk.vercel.ai](https://sdk.vercel.ai/) | ## 🔗 Related GPT4Free Projects

🎁 Projects	⭐ Stars	📚 Forks	🛎 Issues	📬 Pull requests
gpt4free
gpt4free-ts
Free AI API's & Potential Providers List
ChatGPT-Clone
ChatGpt Discord Bot
Nyx-Bot (Discord)
LangChain gpt4free
ChatGpt Telegram Bot
ChatGpt Line Bot
Action Translate Readme
Langchain Document GPT

## 🤝 Contribute #### Create Provider with AI Tool Call in your terminal the `create_provider.py` script: ```bash 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](https://github.com/zukixa/cool-ai-stuff#ai-chat-websites), or find your own provider source! 2. Create a new file in [g4f/Provider](./g4f/Provider) with the name of the Provider 3. Implement a class that extends [BaseProvider](./g4f/Provider/base_provider.py). ```py 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 @classmethod async def create_async_generator( cls, model: str, messages: Messages, proxy: str = None, **kwargs ) -> AsyncResult: yield "" ``` 4. Here, you can adjust the settings, for example, if the website does support streaming, set `supports_stream` to `True`... 5. 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 6. Add the Provider Name in [`g4f/Provider/__init__.py`](./g4f/Provider/__init__.py) ```py from .HogeService import HogeService __all__ = [ HogeService, ] ``` 7. You are done !, test the provider by calling it: ```py 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='') ``` ## 🙌 Contributors A list of the contributors is available [here](https://github.com/xtekky/gpt4free/graphs/contributors) The [`Vercel.py`](https://github.com/xtekky/gpt4free/blob/main/g4f/Provider/Vercel.py) file contains code from [vercel-llm-api](https://github.com/ading2210/vercel-llm-api) by [@ading2210](https://github.com/ading2210), which is licensed under the [GNU GPL v3](https://www.gnu.org/licenses/gpl-3.0.txt) Top 1 Contributor: [@hlohaus](https://github.com/hlohaus) ## ©️ Copyright This program is licensed under the [GNU GPL v3](https://www.gnu.org/licenses/gpl-3.0.txt) ``` 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 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 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 . ``` ## ⭐ Star History

This project is licensed under GNU_GPL_v3.0.

(🔼 Back to top)