Skip to content

Latest commit

 

History

History
81 lines (57 loc) · 3.06 KB

README_en.md

File metadata and controls

81 lines (57 loc) · 3.06 KB
Raw

ChatGPT_API_NoKey

中文|English

Screw the APIKEY! Chat to GPT with NO KEY!

Features

FakeAPI

ChatGPT_API_NoKey is a simple ChatGPT fake API that mimics keyboard and mouse operations, allowing users to interact with GPT without the need for an APIKEY.

Note:

  1. For the first-time use, set HEADLESS=False and manually log in to ChatGPT to save cookies for automatic login in subsequent sessions. The script does not include any other listening functionalities, as assured by the author.
  2. Abusing this API may lead to account issues, so please control the request frequency appropriately.
  3. Currently, only single-thread blocking mode is supported, and concurrent request processing is still under development.

Installation and Usage

  1. Environment Setup

Please install Google Chrome first.

cd <work-dir>
pip install -r .\requirements.txt
  1. Config Configuration

HEADLESS: Whether to run Chrome in headless mode. If set to True, Chrome will run in headless mode without displaying the browser window. If set to False, Chrome will run in normal mode and display the browser window. (Note: There is a small bug in undetected_chromedriver when HEADLESS=True. The solution is provided in example.py.)

HEADER: User-agent header for browser requests.

SERVER_ADDRESS: Server address and port number tuple for the FakeAPI.

PROXY: Proxy used to access ChatGPT. In this configuration, the proxy is set to None, indicating no proxy is used (or the system's default proxy) to access ChatGPT. If you need to use a proxy, you can set the proxy address to the corresponding string, such as "http://127.0.0.1:7890". Please ensure that the proxy is set correctly to ensure proper access to ChatGPT.

  1. Quick Start (example.py includes relevant routines)

run example.py

python ./example.py

Direct Call to FakeAPI

api = FakeAPI(HEADLESS, PROXY, HEADER)
print(api.request("Hello!"))

Using as a Server

Thread(target=SingleThreadServer, args=(HEADLESS, PROXY, HEADER, SERVER_ADDRESS)).start()
time.sleep(15)  # Wait for the server to start (It's better to start the server in another thread)
openai.api_base = "http://localhost:5000"
openai.api_key = "whatever"
completions = openai.ChatCompletion.create(
                    model="whatever",
                    messages="Hello!",
                )
print(completions['choices'][0]['message']['content'])

Future Work

  • Implement multi-threaded concurrent request processing based on Flask.

Acknowledgements

Thanks to the following projects for their valuable contributions to this project:

And other projects that may have been accidentally overlooked :)

Special thanks to the open-source community and all contributors for their contributions to this project.

License

This project is released under the MIT license. For more information, please see the LICENSE file.

Author

Master Yip

Email: [email protected]

GitHub: Master Yip