With the new Firefly custom models, rolling out today in public beta, creators can train a model on their own aesthetic to consistently generate images that match their style. Here are the details.

Adobe is releasing a new capability in Firefly today called .

With it, creators, brands, and users in general can now train the model to follow their own illustration and photographic styles, and have it create new images that consistently preserve details such as stroke weight, color palettes, lighting, and even character features across generations.

According to Adobe, this lets creators “generate new ideas aligned to your aesthetic, reuse the model across projects, briefs and campaigns and produce at scale without losing what makes your work distinctive.”

The beta feature is , and each training currently costs 500 credits from the user’s monthly generative balance.

There are currently three different raining methods. Here’s Adobe

• Photography style: Train on a unique look — lighting, color, and mood — to generate new shots with the same visual feel.
• Illustration style: Create fresh illustrations that carry your signature look.
• Characters: Bring character to life consistently across scenes and stories.

Users can send between 10 and 30 JPG or PNG images with a minimum resolution of 1024×1024, and they should have a consistent aspect ratio (up to 16:9 landscape or 9:16 portrait).

The company’s for achieving the best results include uploading images with a consistent style and color palette, and avoiding low-resolution or blurry images.

Adobe has also published a to training Firefly custom models.

Today, Adobe is also opening a private beta for Project Moonlight, a feature with a conversational interface that lets users describe what they want to accomplish in a turn-by-turn chat, rather than a simple prompt box.

These agentic AI assistants will be available “across Adobe’s products, including Photoshop, Express and Acrobat,” and will work autonomously to “execute that vision using Adobe’s iconic tools – taking real actions you can refine, adjust and build on.”

To learn more about Project Moonlight and join the private beta’s waitlist, .

Finally, Adobe also announced today that Adobe Firefly now includes access to 30+ top AI image and video models, including Google’s Nano Banana 2 and Veo 3.1, Runway’s Gen-4.5, and Adobe’s Firefly Image Model 5.

This is particularly relevant, as Adobe is still offering unlimited video and image generation through April 22, and you can learn more about current Firefly promotions .