Google AI Studio

I recently added a "Built by Aaron Francis" section with my picture to Soloterm.com. I wanted to show how I created that image using AI.

First, I started with a regular photo of me with my twins and removed the background using Photoshop (though any background removal tool would work). Then, I used Google's AI Studio, specifically Nano Banana Pro, after figuring out how to set up the API key. I set the resolution to 4K and started experimenting with prompts.

Initially, my attempts were terrible. I wanted to style the image with a crosshatch pattern to match the website's aesthetic. I tried different prompts, focusing on restyling the image to blend with website screenshots. I found that generative AI can be unpredictable.

One of my attempts actually turned out okay, even though the AI added a fake transparent background. I kept refining my prompts, realizing I was going off track at one point and decided to start over. Eventually, I got the image I wanted, but I still had to remove the background again in Photoshop.

My successful prompt involved updating the image to have a blue crosshatch that would fit the website. I showed some failed attempts and one that was decent, noting it made me look skinnier.

After finding a style I liked, I'd take the image to Gemini or ChatGPT and ask it to describe the style in a prompt format so I could reuse it on other images. Midjourney is also good for copying styles. Overall, Google AI Studio is useful for quick image work, but Midjourney is better for heavy image manipulation.