taranjeet/unofficial-chatgpt-api

This repo is unofficial ChatGPT api. It is based on Daniel Gross's WhatsApp GPT

ChatGPT api

• It uses playwright and chromium to open browser and parse html.
• It is an unoffical api for development purpose only.

How to install

• Make sure that python and virual environment is installed.

• Create a new virtual environment

# one time
virtualenv -p $(which python3) pyenv

# everytime you want to run the server
source pyenv/bin/activate

• Now install the requirements

pip install -r requirements.txt

• If you are installing playwright for the first time, it will ask you to run this command for one time only.

playwright install

• Now run the server

python server.py

• The server runs at port 5001. If you want to change, you can change it in server.py

Api Documentation

• There is a single end point only. It is available at /chat

curl -XGET http://localhost:5001/chat?q=Write%20a%20python%20program%20to%20reverse%20a%20list

Updates

• [8 Dec 2022]: Updated parsing logic (credits @CoolLoong)

Credit

• All the credit for this script goes to Daniel Gross's whatsapp gpt package. I have just taken the script as an individual file and added documentation for how to install and run it.

Disclaimer

Please note that this project is a personal undertaking and not an official OpenAI product. It is not affiliated with OpenAI in any way, and should not be mistaken as such.

unofficial-chatgpt-api's Issues

Write API interface for native chatgpt model

Incorrect usage of the term 'API' in project description

This project is not an API, but rather a project that utilizes a Chrome-based automation library to open the chat.openai.com website and retrieve the output from that website. This output is not an API, but rather the direct output from the OpenAI website. Therefore, it is not appropriate to refer to this project as an API. Instead, it should be referred to as a project that utilizes the output from the OpenAI website in order to achieve its intended purpose.

This no longer works as chatGPT had made changes to their frontend

well, the queries which the script is looking for in the chatGPT page no longer exists

Getting this right after I login and hit enter

░▒▓ ~/g/chatgpt-api │ on main ▓▒░ node:events:491 ░▒▓ ✔ │ took 30s │ pyenv Py │ at 11:49:40 ▓▒░
throw er; // Unhandled 'error' event
^

Error: write EPIPE
at WriteWrap.onWriteComplete [as oncomplete] (node:internal/stream_base_commons:94:16)
Emitted 'error' event on Socket instance at:
at emitErrorNT (node:internal/streams/destroy:157:8)
at emitErrorCloseNT (node:internal/streams/destroy:122:3)
at processTicksAndRejections (node:internal/process/task_queues:83:21) {
errno: -32,
code: 'EPIPE',
syscall: 'write'
}
Can you help me with this

Improve response time

Hi,
thanks for the great repo!
Currently, the python scripts waits 10 seconds before returning the response. This can lead to errors for example if the result takes longer.

You can take this code (or something similar) that looks for the loading indicator and immediately returns the result when the indicator disappears.

...    
send_message(message)

  # Check if an answer was received
    while True:
        # Check loading indicator
        loading = PAGE.query_selector_all(
            "div[class*='result-streaming']")
        if len(loading) == 0:
            response = get_last_message()
            break

print("Response: ", response)
...

Thanks for the great work!

too many errors

如何使用chatgpt模型写一个本地应用

how to write a local application with chatgpt model

Didn't work for me, followed the steps, getting playwright._impl._api_types.Error: value: expected string, got undefined

playwright._impl._api_types.Error: value: expected string, got undefined

为本地chatgpt模型写API接口

为GPT-chitchat模型写后端接口

I can't get passed the verify you are a human bottom. Any suggestions?