After the AI video tool I decided to experiment a bit with a pivot towards ads. So I just added the option to use input images of your products so you can just add like for example the image of a sneaker and it would generate an ad for that sneaker. It works but the tech is not entirely there yet. The image gen model (gpt-image-1) kinda works. But the video generation from that initial frame fucks it up.

Here’s an example of a Nike sneaker ad:

And a Gucci bag: