Adobe Firefly Adds Prompt-Based Video Editing and AI Models

Adobe is expanding the capabilities of its Firefly platform, rolling out a comprehensive video editor that allows for granular, prompt-based adjustments. The update, which follows a private beta phase initiated in October, is now available to all users, marking a shift from simple generation to complex video manipulation.

Advanced Editing Controls

Previously, Firefly users were restricted to generating video from scratch; any desired changes required a full regeneration of the clip. The new editor introduces a timeline-based interface, enabling direct manipulation of frames, sound, and visual elements.

By integrating Runway’s Aleph model, the platform now accepts specific directives, such as adjusting contrast, modifying weather conditions like “change the sky to overcast,” or executing camera movements like “zoom in slightly on the main subject.” Furthermore, Adobe’s proprietary Firefly Video model now supports reference-based editing, allowing users to upload a starting frame and a reference video to replicate specific camera angles.

Expanded Model Integration

Adobe is also diversifying its ecosystem by incorporating third-party AI models to enhance creative workflows:

• Topaz Labs Astra: Now available for high-quality video upscaling to 1080p or 4K.
• Black Forest Labs FLUX.2: An image generation model now accessible across Firefly platforms, with integration for Adobe Express users scheduled for January.

Incentives for Subscribers

To encourage adoption of these tools, Adobe is offering unlimited generations for specific subscription tiers—including Pro, Premium, 7,000-credit, and 50,000-credit plans—within the Firefly app until January 15. The company continues to evolve its AI suite, having introduced subscription models in February followed by the launch of dedicated web and mobile applications throughout the year.