There's two steps. First, you set up the questions you want to ask about a website. Lets say you are scraping Github repos. Your questions might be "What is the language used in this repo" and "How many stars does it have".

Then, you visit that website and click a button in the Chrome extension. FetchFox sends a request to GPT-4 of the innerHTML of the website, and it gets back the answers. If it can't find the answer, it says so.

I built this tool because I do a lot of webscraping, and I often run into a few problems: websites with complicated obfuscated HTML, difficult to parse API responses on SPA's, and of course IP blocking and anti-scraping measures. This Chrome extension does an end run around all of those.

Right now you can only use it for small scale scraping jobs, but I'd love to get the community's feedback on how this can be made more useful.