Browser-gpt-api

Interact with ChatGPT via an API running on a local server using Selenium under the hood.

What is it?

A python tool to help you send prompts to ChatGPT webpage and get response via API calls. With Browser-GPT you can get infinite amount of API calls without worrying about any credits. This works by automating the process of sending a prompt to ChatGPT and getting the response back using Selenium.

How to Install & run:

This project uses Rye Package manager

1. Clone the repository:

   git clone https://github.com/vertexcover-io/browser-gpt-api.git
   cd src/browser_gpt_api
   

2. Install project dependencies using Rye:

   rye sync
   

3. Create a copy of the environment variable template file and edit it according to your configuration:

   cp .env.template .env
   

4. Download Chrome Driver based on your chrome version and OS.

5. Add the location of chromedriver.exe file as a System variable in Environmental variables.

6. In the main directory run setup.py:

   python setup.py
   

7. To setup your local-server run server.py:

   python server.py
   

8. Now you can access ChatGPT via the API endpoint : http://127.0.0.1:8000/chat/completion

Stack

• Rye
• Selenium
• FastAPI
• uvicorn

Pre-Commit

Set up pre-commit hooks to automatically check your code for linting and formatting issues before the commit. Run following command to install pre-commit hooks:

rye run pre-commit install

Format and Lint

This project uses Ruff for code formatting and linting. Use following commands to format and lint your code:

• Check for issues:

ruff check 

• Format your code:

ruff format

For more information refer to the Ruff Documentation.