Autocoder is a frontend to OpenAI's GPT-3, made with access to the beta generously provided by OpenAI at the recent Scale Transform Conference.
It takes a description of a webpage and renders the page automatically. It is currently trained on simple html elements and uses the ada and davinci semantic and completion models respectively. The API endpoint used is /answers.
Inspired by the demo of Sharif Shameem.
It is written in Python, with a Bootstrap frontend. The API was implemented in Flask and there are a series of unit tests in test_suite.py.
Install requirements
pip install -r requirements.txtStart the server:
python app.pyNavigate to localhost:5000 in your browser, enter your prompt and click Submit.
Please note: you will need an API key from OpenAI to construct the post request - save it in a file called config.py as OPENAI_API_KEY.
Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.
Please make sure to update tests as appropriate.