Adobe has announced a major expansion of its Firefly platform, introducing new AI capabilities, custom model training, and enhanced video and image creation tools aimed at streamlining creative workflows and enabling more personalized content generation.
The update positions Firefly as an all-in-one creative AI studio, integrating more than 30 industry-leading models from companies including Adobe, Google, OpenAI, and Runway. The platform now allows users to generate, refine, and edit creative outputs within a single environment, reducing friction between ideation and production.
A central feature of the release is the introduction of Firefly custom models, now available in public beta. These models allow users to train AI on their own images to replicate specific artistic styles, characters, or photographic aesthetics. By uploading their own assets, creators can build reusable models that preserve visual consistency across projects, including elements such as color palettes, lighting, and design details.
Adobe emphasized that these custom models are particularly useful for illustration styles, character consistency, and repeatable photographic looks. Once trained, the models can be reused across campaigns and creative workflows, helping teams scale content production without losing brand identity. The company also noted that these models are private by default, ensuring that user-generated content remains proprietary.
In addition to custom models, Firefly now offers expanded creative flexibility through access to a broad range of AI models, each optimized for different use cases such as cinematic motion, photorealism, or stylized design. Users can generate content with one model, refine it with another, and continue editing using Adobe’s native tools within the same interface.
Adobe also introduced new editing capabilities designed to accelerate production workflows. Features like Quick Cut enable users to transform raw video footage into structured edits quickly, while enhanced image editing tools allow for object manipulation, scene extension, and fine-tuning of generated visuals. These tools are fully integrated with the generation process, enabling a seamless transition from concept to finished output.
Another key development is the shift from prompt-based creation to conversational AI workflows. Adobe is rolling out agentic AI assistants across products such as Photoshop, Express, and Acrobat. These assistants allow users to describe creative goals in natural language, with the system executing tasks and enabling iterative refinement through dialogue.
As part of this evolution, Adobe is expanding access to Project Moonlight, a conversational AI interface currently in private beta. The system is designed to work across Adobe applications, helping users move from initial concept to final output more efficiently while maintaining creative control.
The company framed these updates as part of a broader transformation in how creative work is produced, moving toward more interactive, AI-assisted workflows that combine generation, editing, and collaboration in a unified environment.
