Adobe’s AI image generator can now be trained on your own art
TL;DR
Adobe has launched Firefly Custom Models in public beta, letting users upload their own images so the AI learns and replicates specific artistic styles and aesthetics.
Key Points
- The feature targets creators and brands that need large volumes of visually consistent content – including character designs, illustrations, and photography.
- Once trained, a custom model acts as a reusable foundation across projects, eliminating the need to start from scratch each time.
- The tool integrates into Adobe's existing Creative Cloud ecosystem and is designed to speed up high-volume visual production workflows for teams.
Nauti's Take
Adobe is doing the right thing here: instead of generic prompt engineering, finally a system that adapts to your own visual language. The clever part is not the underlying tech – LoRA-style fine-tuning on custom data has existed for years – but the packaging into a legally clean, Creative-Cloud-native product.
Agencies and brands that held back from Midjourney and others due to IP uncertainty now have a serious alternative worth evaluating. The real question is output quality – public beta still means work in progress.
Video
Sources