The GPT-V-on-Web uses GPT-4V(ision) and Python to build an interactive web agent that allows users to navigate and interact with websites conversationally using simple text prompts.
It can click links, fill out forms, search pages, and more based on natural language instructions. This enables hands-free web browsing and automation of repetitive online tasks.
How to use it:
1. Install GPT-V-on-Web and chromedriver (Mac users) packages.
# For Mac users brew install chromedriver # OR brew upgrade chromedriver # Install pip install git+https://github.com/Jiayi-Pan/GPT-V-on-Web.git
2. Type webai and press enter. GPT-V-on-Web will start and take over your web interactions.
3. To start on a particular website, type webai --start_link "https://www.scriptbyai.com" in the terminal. The AI will open the specified link and be ready for your prompts.
