Skip to content

Latest commit

 

History

History
223 lines (160 loc) · 5.39 KB

README.md

File metadata and controls

223 lines (160 loc) · 5.39 KB
Raw

ChatGPT

PyPi Downloads

Reverse Engineered ChatGPT API by OpenAI. Extensible for chatbots etc.

Support my work

Make a pull request and fix my bad code.

Installation

pip3 install --upgrade revChatGPT

V1 Standard ChatGPT

Uses chat.openai.com

  • Free
  • Sunset in a few weeks due to high costs

Proxy server Rate limit: 10 requests per 10 seconds (per IP) - This is due to high abuse. I'm trying to keep it at a sustainable level so I don't run out of funds too quickly.

OpenAI rate limit: 50 requests per hour on free accounts. You can get around it with multi-account cycling

Plus accounts has around 150 requests per hour rate limit

Configuration

  1. Create account on OpenAI's ChatGPT
  2. Save your email and password

Authentication method: (Choose 1)

- Email/Password

Not supported for Google/Microsoft accounts

{
  "email": "email",
  "password": "your password"
}

- Session token

Comes from cookies on chat.openai.com as "__Secure-next-auth.session-token"

{
  "session_token": "..."
}

- Access token

https://chat.openai.com/api/auth/session

{
  "access_token": "<access_token>"
}

- Optional configuration:

{
  "conversation_id": "UUID...",
  "parent_id": "UUID...",
  "proxy": "...",
  "paid": false
}
  1. Save this as $HOME/.config/revChatGPT/config.json
  2. If you are using Windows, you will need to create an environment variable named HOME and set it to your home profile for the script to be able to locate the config.json file.

Usage

Command line

python3 -m revChatGPT.V1

        ChatGPT - A command-line interface to OpenAI's ChatGPT (https://chat.openai.com/chat)
        Repo: github.com/acheong08/ChatGPT

Type '!help' to show a full list of commands

Logging in...

You:
(Press Esc followed by Enter to finish)

The command line interface supports multi-line inputs and allows navigation using arrow keys. Besides, you can also edit history inputs by arrow keys when the prompt is empty. It also completes your input if it finds matched previous prompts. To finish input, press Esc and then Enter as solely Enter itself is used for creating new line in multi-line mode.

Set the environment variable NO_COLOR to true to disable color output.

Developer API

Basic example (streamed):

from revChatGPT.V1 import Chatbot

chatbot = Chatbot(config={
  "email": "<your email>",
  "password": "<your password>"
})

print("Chatbot: ")
prev_text = ""
for data in chatbot.ask(
    "Hello world",
):
    message = data["message"][len(prev_text) :]
    print(message, end="", flush=True)
    prev_text = data["message"]
print()

Basic example (single result):

from revChatGPT.V1 import Chatbot

chatbot = Chatbot(config={
  "email": "<your email>",
  "password": "<your password>"
})

prompt = "how many beaches does portugal have?"
response = ""

for data in chatbot.ask(
  prompt
):
    response = data["message"]

print(response)

All API methods

Refer to the wiki for advanced developer usage.

V3 Official Chat API

Recently released by OpenAI

  • Paid

Get API key from https://platform.openai.com/account/api-keys

Command line

python3 -m revChatGPT.V3 --api_key <api_key>

 $ python3 -m revChatGPT.V3 -h

    ChatGPT - Official ChatGPT API
    Repo: github.com/acheong08/ChatGPT

Type '!help' to show a full list of commands
Press Esc followed by Enter or Alt+Enter to send a message.

usage: V3.py [-h] --api_key API_KEY [--temperature TEMPERATURE] [--no_stream] [--base_prompt BASE_PROMPT]
             [--proxy PROXY] [--top_p TOP_P] [--reply_count REPLY_COUNT] [--enable-internet]

options:
  -h, --help            show this help message and exit
  --api_key API_KEY     OpenAI API key
  --temperature TEMPERATURE
                        Temperature for response
  --no_stream           Disable streaming
  --base_prompt BASE_PROMPT
                        Base prompt for chatbot
  --proxy PROXY         Proxy address
  --top_p TOP_P         Top p for response
  --reply_count REPLY_COUNT
                        Number of replies for each prompt
  --enable_internet     Allow ChatGPT to access the internet (beta)

Developer API

Basic example

from revChatGPT.V3 import Chatbot
chatbot = Chatbot(api_key="<api_key>")
chatbot.ask("Hello world")

Streaming example

from revChatGPT.V3 import Chatbot
chatbot = Chatbot(api_key="<api_key>")
for data in chatbot.ask("Hello world"):
    print(data, end="", flush=True)

Awesome ChatGPT

My list

If you have a cool project you want added to the list, open an issue.

Disclaimers

This is not an official OpenAI product. This is a personal project and is not affiliated with OpenAI in any way. Don't sue me.

Credits