Adobe has expanded its Firefly platform with new AI capabilities, including custom models, additional third-party integrations and enhanced video and image creation tools.
The update has introduced Firefly custom models in public beta, allowing users to train AI models using their own images. These models have been designed to replicate specific styles, characters or photographic aesthetics, enabling consistent output across projects.
The platform has supported use cases such as illustration styles, character consistency and repeated photographic looks. It has enabled users to reuse trained models across campaigns and creative workflows while maintaining visual consistency. The models have remained private by default.
Adobe Firefly has also integrated more than 30 AI models from multiple providers, including those from Adobe, Google, OpenAI, Runway and Kling. Users have been able to generate content using one model and refine it using another within the same environment.
The company has introduced new tools for video and image editing. Features such as Quick Cut have enabled users to convert raw footage into structured edits, while expanded image editing tools have supported object removal, scene extension and visual adjustments.
Adobe has also introduced conversational AI workflows across its products, including Adobe Photoshop, Adobe Express and Adobe Acrobat. These tools have allowed users to create and edit content using chat-based inputs.
Additionally, the company has expanded access to Project Moonlight, a conversational AI interface currently in private beta. The feature has been designed to help users move from concept to execution using agentic AI workflows.
