Adobe Firefly adds AI video editing and third-party models
▼ Summary
– Adobe is updating its Firefly AI app with a new video editor that enables precise, prompt-based edits to elements like color and camera angles, moving beyond full-clip regeneration.
– The update introduces new third-party AI models, including Runway’s Aleph for specific edits, Topaz Labs’ Astra for video upscaling, and Black Forest Labs’ FLUX.2 for image generation.
– A new timeline view in the editor allows users to easily adjust frames, sounds, and other video characteristics.
– To attract users amidst competition, Adobe is offering unlimited generations from its image and video models to certain subscription plan holders until January 15.
– This is part of a series of major updates to Firefly this year, including new subscription tiers, web and mobile apps, and increased third-party model integration.
Adobe has significantly upgraded its Firefly AI application, introducing a powerful new video editor and expanding its creative toolkit with several third-party generative models. This move transforms Firefly from a basic generation tool into a more sophisticated editing suite, directly addressing a key user need for greater control over AI-generated video content.
Previously, creating a video with Firefly was an all-or-nothing proposition. If any element of the generated clip was unsatisfactory, users had to start the entire process over with a new text prompt. The new video editor changes this dynamic entirely, allowing for precise, prompt-based modifications. Users can now instruct the AI to alter specific elements like colors, objects, or camera angles without regenerating the whole video. A new timeline interface makes it simple to adjust individual frames, sounds, and other characteristics, bringing a familiar editing workflow to AI video creation.
This editor, first shown in a private beta last October, is now available to all Firefly users. It leverages multiple AI models to execute detailed edits. For instance, using Runway’s Aleph model, a creator could type a command such as “Change the sky to overcast and lower the contrast” to instantly alter the mood of a scene. Adobe’s own Firefly Video model enables more complex tasks, like uploading a starting frame and a reference clip with a specific camera movement, then directing the AI to apply that same motion to the project.
Beyond editing, Adobe is broadening the generative capabilities within Firefly by integrating models from other leading AI companies. Users can now employ Topaz Labs’ Astra model to upscale videos to crisp 1080p or 4K resolution. Furthermore, the popular FLUX.2 image generation model from Black Forest Labs is being added to the platform. A new collaborative boards feature is also included, facilitating team-based creative projects. Adobe states that FLUX.2 is available immediately in Firefly across platforms, with integration into Adobe Express coming in January.
These substantial updates arrive amidst fierce competition in the AI media generation space. To encourage adoption and deep engagement with its enhanced app, Adobe is offering a generous incentive. Subscribers to various Firefly Pro, Premium, and high-credit plans will receive unlimited generations from all image models and the Adobe Firefly Video Model until January 15. This follows a year of rapid evolution for Firefly, which has seen the launch of tiered subscriptions, a dedicated web app, mobile applications, and steadily growing support for external AI models.
(Source: TechCrunch)
