My son was shopping for middle school supplieswhen he ran into a problem I assumed Amazon had solved years ago.
His best friend didn’t have the school supply list, so my son wanted to send him the contents of his own Amazon cart. The idea was simple: compare what they had picked and make sure they were buying the same binders, notebooks and other supplies.
Except there was no share button. Well, there is, but only for businesses using Amazon business. The only option for consumers is to share Lists, but that would have required my son to create a new list and move everything over after he had already filled his cart. It’s possible, but hardly intuitive. He didn’t want to create a registry. He wanted to send his friend the cart that was already sitting in front of him.
Amazon does have an official Share Cart feature, but it’s designed for Amazon Business accounts. That felt like a strangely obvious feature for Amazon to leave out. So I decided to see whether I could build it myself.
I’m always looking for a reason to build something, but I really wondered if it would be possible to build something like this. For those who aren’t developers, vibe codinghas made it incredibly easy to see a problem and build a solution using natural language.
In this case, I described the problem to ChatGPT: I wanted a button that could read the products and quantities in an Amazon cart, turn them into a private link and let someone else open that link to see the same supplies.
The result was CartShare, a small website paired with a Chrome extension.
The proposed flow was exactly what my son had asked for. Open an Amazon cart, click the CartShare extension and copy the resulting link. The friend receiving it could check off anything they already owned, open individual products on Amazon or select the supplies they wanted.
The shared cart itself didn’t require an account or database. Product identifiers, titles and quantities were encoded inside the link, which meant CartShare didn’t need anyone’s Amazon password, delivery address or payment information.
ChatGPT built the interface, packaged the Chrome extension and published the site. It looked like a real product, complete with a name, privacy language and a large button promising to copy the selected products into the recipient’s Amazon cart.
Does it actually work?
Yes…and no. The sharing page worked. CartShare could display a set of products, preserve their quantities and create a link that could be sent to someone else.
The one-click Amazon transfer did not. The prototype attempted to use an Amazon cart URL containing the product identifiers and quantities. When I tested it, Amazon returned a 500 error and said it didn’t support the session or client.
That meant the most satisfying part of the idea, pressing one button and watching someone else’s cart populate, couldn’t be treated as finished. The Chrome extension that reads the original cart still needs to be tested against a real Amazon cart as well, since Amazon can change the structure of its pages at any time.
I went back to the CartShare site and changed it. The transfer button was removed and replaced with a clear status notice explaining that the project worked only partially. Recipients could compare products and open them individually, but CartShare could not yet promise a complete cart-to-cart transfer.
It was a useful reminder of how deceptive vibe coding can be, even when nobody is trying to deceive anyone. AI can produce a polished interface before the underlying behavior has been properly tested. The buttons, colors and copy make the project feel finished, but a convincing website is not proof that the feature behind it works especially when you put it up against a big corporation like Amazon.
Bottom line
Could CartShare eventually work? Probably, although it may require both people to install the extension.
Instead of relying on Amazon’s rejected cart URL, a future version could open the recipient’s Amazon cart and add each selected product through the same visible controls a shopper would use. That would be slower and more fragile because the extension would need to keep up with changes to Amazon’s website. It would still avoid sharing passwords or payment information.
The other option is to stop short of automatically filling the recipient’s cart. Even a clean, shareable page showing the exact products and quantities is more convenient than asking someone to rebuild a cart from screenshots or create an Amazon List after the shopping is already done.
CartShare isn’t ready to replace Amazon’s missing share button. It is a public prototype that shows how the feature could work and, perhaps more importantly, where the simple idea runs into Amazon’s technical boundaries.
My son’s request took seconds to explain. Vibe coding turned it into a website and Chrome extension without requiring me to become a developer first. Getting from a convincing prototype to a dependable product is where the harder work begins but overall, I feel like this was a win. I personally enjoy seeing how things work, even if they don’t fully come to fruition.
- Google Gemini just made it easy to remove watermarks from AI images — here’s how to do it
- I don’t use AI to write my emails — here’s how I use it to manage my inbox instead
- 5 features that give Claude the edge over ChatGPT
