ChatGPT Wrapper is an open-source unofficial Python API and CLI that lets you interact with ChatGPT.
🤖 Programmable ChatGPT. The ChatGPT Wrapper lets you use the powerful ChatGPT bot in your Python scripts or on the command line, making it easy to leverage its functionality into your projects.
💬 Runs in Shell. You can call and interact with ChatGPT in the terminal
Run an interactive CLI in the terminal:
Or just get a quick response for one question:
See below for details on using ChatGPT as an API from Python.
To use this repository, you need setuptools
installed. You can install it using pip install setuptools
.
- Install the latest version of this software directly from github with pip:
pip install git+https://github.com/mmabrouk/chatgpt-wrapper
- Install a browser in playwright (if you haven't already). The program will use firefox by default.
playwright install firefox
- Start up the program in
install
mode. This opens up a browser window. Log in to ChatGPT in the browser window, then stop the program.
chatgpt install
- Restart the program without the
install
parameter to begin using it.
Note: For more help, this Blog post provides a visual step-by-step guide for installing this library.
In addition to directly prompting ChatGPT, The shell provides the following commands:
!new
allows to start a new conversation!nav
lets you navigate to a past point in the conversation. Example:!nav 2
!stream
toggles between streaming mode (streams the raw response from ChatGPT) and markdown rendering (which cannot stream).!read
begins reading multi-line input. Usefull when copy-pasting!file
sends a prompt read from the named file. Example:file myprompt.txt
!log
enables logging to a file. Example:!log mylog.txt
to enable, or!log
to disable.!context
lets you load old contexts from the log (previous sessions). It takes one parameter; a context string from logs.!session
refreshes your session information. This can resolve errors under certain scenarios.
To use the ChatGPT
class as an API for talking to ChatGPT, create an instance of the class and use the ask
method to send a message to OpenAI and receive the response. For example:
from chatgpt_wrapper import ChatGPT
bot = ChatGPT()
response = bot.ask("Hello, world!")
print(response) # prints the response from chatGPT
The say method takes a string argument representing the message to send to ChatGPT, and returns a string representing the response received from ChatGPT.
You may also stream the response as it comes in from ChatGPT in chunks using the ask_stream
generator.
We welcome contributions to ChatGPT Wrapper! If you have an idea for a new feature or have found a bug, please open an issue on the GitHub repository.
This project is licensed under the MIT License - see the LICENSE file for details.
- 18/01/2022: v0.3.8 - Commands now are run only using !. For instance to enable read mode (for copy-paste and long prompts) you need to write now
!read
instead ofread
. This is to avoid conflicts with the chatgpt prompts. Fixed timeout issue. - 17/01/2022: v0.3.7 - Added timeout to
ask
method to prevent hanging. Fixed return to terminal breakdown. Streaming output now is activated by default.
- This project is a modification from Taranjeet code which is a modification of Daniel Gross code.