AI generated. Prompt: A product photograph of a beautiful bracelet made from hammered sterling silver with a labradorite stone in the center

Let’s Make a Bracelet: A Comparison of AI Image Generators

My hands-on experience with AI image generators is limited, so I don’t know which one is most likely to create achievable designs. I decided to try out a few different options and put the results side by side. I chose to attempt to generate an image of a bracelet I made in a metalsmithing class I took a few years ago. The prompt was, “A product photograph of a beautiful bracelet made from hammered sterling silver with a labradorite stone in the center.”


I have the most experience with Open AI’s DALL-E, so that’s where I started. Here are the four images DALL-E generated from the prompt:

They look nothing like the original, simple hammered bracelet I envisioned, but they are certainly unique. While each has its own charm, I am most fond of the image on the right. I like the wave-like curves of what could be handmade chain, and the simple presentation of a round stone in the center.

Elementor AI

My next stop was Elementor AI. Elementor is the page builder I use for Eliza G. Designs, and they’ve recently started offering AI enhancements as part of the program. I’ve been impressed by what I’ve seen so far, so I eagerly anticipated the results of my prompt.

Like DALL-E, Elementor AI’s image generator creates four variations from which to choose, but while DALL-E lets you keep them all, Elementor AI requires you to choose just one image. The rest are forever lost, pixels in space. I could have done a screen capture, but that felt cumbersome and, truth be told, the initial four results were remarkably similar in appearance, as if the generator had created four different shots of the same “product.” They’re beautiful pieces, and what I wouldn’t give to find a labradorite that blue!

I re-generated the prompt and got similar results, but this time the generator included models in some of the images. Again, the jewelry was similar, but I like the piece the model is wearing. The relatively large, somewhat irregular stone combined with what appears to be a sort of braided silver band is intriguing.


Next I decided to give Gencraft a go. I wasn’t familiar with it, but it had a free option (with registration), so I input my prompt and waited with no real expectations. I was pleasantly surprised at the results.

Now we’re getting somewhere! These pieces have a wonderful handmade look, and the irregular stone shapes give them an organic feel. Because they seem to follow some sort of real-world logic, it would be possible for a skilled artisan to almost perfectly re-create these looks. I’ll need a lot more experience before taking on a creation of this complexity, but the possibilities excite me.


I was also unfamiliar with starryai, which I found cumbersome to use, and many basic options required a pro upgrade. The images bear little resemblance to the prompt, but the designs are rather striking.


Canva’s generator had some surprisingly commercial ideas that could be easily recreated. There’s a certain cheapness to the images that makes them feel less artistic, but the jewelry depicted could be made by a relative beginner. Except for that third image—what is that?


Fotor likes beads. I like beads, too. In fact, I could practically make designs similar to these from beads I have in my stash. I’m enjoying imagining the piece on the right with a heart-shaped bead at the forefront.


After “designing” so many bracelets, I had a hankering to make something in real life, so I decided to bring the day’s experiment to a close. I’ll audition more image generators in the near future; new ones seem to be popping up every day!

Browsing through the generated photos, I have a hard time picking a clear winner. Every AI image generator created at least a couple of usable, intriguing designs. DALL-E’s designs seem the most surreal, while Canva’s are almost too realistic. I will probably stick with Elementor AI for the most part, because it is so exceedingly simple to pop an image onto the website, and it seems to create fairly realistic objects, especially when they are displayed on a model. But I will definitely continue to use a variety of AI image generators, depending upon the results I hope to achieve.

I hope you’ve enjoyed this tour of AI image generators, and find the creativity and variety as fascinating as I do. Come back for more and stick around to see what results from all the possibilities.

Leave a Reply

Your email address will not be published. Required fields are marked *