Exploring Adobe Firefly: Image Generation and Editing Essentials

After working with Photoshop, Illustrator, and InDesign, it’s time to review Adobe Firefly. To get started, open the Creative Cloud app. On a Mac, you can click the Creative Cloud icon in the menu bar. On Windows, click the Windows icon in the taskbar, search for Creative Cloud, and open it. Once the app is open, select the Adobe Firefly icon.

Firefly will automatically log you into your Adobe account. In the Firefly interface, you’ll see several tools and options. At the center is the prompt area, where you describe what you want to generate. At the top of the prompt section, you can choose whether you want to generate an image or a video, and you can also select the model you want to use.

Adobe prominently promotes Firefly as a commercially safe model. Below that, you’ll see a long list of additional models you can choose from. This is similar to Photoshop, but the list of available models is much larger here. You can also choose different aspect ratios for your output.

If you click More, Firefly takes you to a settings panel with expanded options. I prefer using this view because it shows all available controls in one place. While NanoBanana is very popular right now, I’ll start by using Firefly itself so we can compare results later.

I’ll begin with Firefly Image 4, which uses fewer credits and is useful for testing. Firefly Image 5 is also available, but I’ll stick with version 4 for now. I’ll set the aspect ratio to 16:9 widescreen and choose Photo as the content type. There’s also a Visual Intensity slider—moving it to the right makes the image more stylized and AI-like, almost like a drawing or 3D render. I’ll leave it at the default setting.

Firefly also allows you to add reference images, upload your own styles, or select from built-in styles and effects. You can adjust color and tone, lighting, and even camera angle, which is especially useful if you’re aiming for a specific perspective. For this example, I’ll leave all of these options off and keep the settings simple.

Now I’ll describe the image I want to generate. I’ll enter: a group of friends drinking coffee in a meeting. I like using group scenes because they help reveal how well the model handles faces and hands, which are often challenging.

Firefly generates four images. From the top menu, I can download or review each result. When I open one of the images, it’s clear that this looks more like an illustration than a photorealistic photo. While it looks good stylistically, it’s not fully realistic.

Some of the generated images are very similar to each other. A few look more realistic than others, and in some cases, the hands look decent, while in others, they start to break down. This variation is common. Still, the results are generally usable depending on your goal.

If I want to refine an image, I can add a new prompt to edit it. For example, I can type add a cat and generate again. Firefly analyzes the entire image and incorporates the new element. This is useful if you want to build on an existing result, change colors, or add objects. Even with a simple prompt, the results are quite good, though hands and fine details may still show issues.

You can also browse the gallery to see examples of what Firefly can generate. The overall visual quality is impressive.

In addition to generating new images, Firefly lets you edit existing images. I’ll click New on the left and upload an image from my computer, this time, an image of the Taj Mahal. Once the image is loaded, I can describe what I want to add. For example, I’ll type add a school bus and bikes and click Create.

Firefly interprets the image and integrates the new elements into the scene. The result looks fairly realistic, even with some background blur. Once I’m satisfied, I can download the image.

I’ll open both the original and the Firefly-generated version in Photoshop to compare them. The original image is 6000 × 3000 pixels, while the Firefly version is smaller, around 2000 pixels wide. Although the resolution is lower, it’s still usable for smaller prints or layouts. The integration quality is quite good considering the size.

Firefly also keeps a history of your generated images. If I select one from the gallery, I can see that it was created with Firefly. I can reuse the same prompt or modify it. For example, I’ll generate a new image with the prompt: a group of corporate people in the future drinking coffee in a modern office meeting.

I’ll generate this using Firefly Image 5 first. The results look solid, with improved hand structure and more realism. Next, I’ll switch the model to Gemini NanoBanana and generate the same prompt. You’ll notice a Google logo appear on these images, indicating the model source.

The NanoBanana results introduce more futuristic elements, screens, lighting, and background details, and the hands appear stronger and more consistent. This is one reason Adobe integrates multiple models: each has different strengths.

You can return to the settings at any time by clicking More or the Firefly logo. Firefly gives you the flexibility to experiment with different models, but it’s important to keep track of your credits. Generating large or multiple images, especially with third-party models, can consume credits quickly. You can check and manage your credits directly from the Firefly interface.

Overall, Adobe Firefly is a powerful tool for image generation and editing. The key is to experiment with different models, prompts, and settings to understand which approach works best for your needs. Now it’s your turn to explore Firefly and start testing what it can do.