[Need help] Issue with the API's results

Hi everyone,
I come to you because I’m stuck with something I’m working on.

So, I created a master prompt to roast a landing page (ChatGPT has to browse the url), and when I use it in ChatGPT 4, it works wonderfully. ChatGPT answer is perfect.

Now, when I do the same but through the API, it’s like the url has not been browsed at all. Only generic content comes out of it. I’m referring to chatgpt 4 model though.

And I’m stuck trying to make it work. I don’t get why. Anyone can help me understand this behavior? Thanks!

Hi @juliand and welcome!

The short answer is that the API does not inherently have the browsing capabilities just like ChatGPT. However, it’s possible to build this in. There are a few different options available. I am including below links to some other posts that discuss these and that you may find helpful.