Adobe Ramps Up Firefly Video Tools Including New Firefly Video Editor

Adobe has expanded its Firefly generative AI platform with new tools aimed squarely at video creators.

Adobe has added a whole load of new video-first features into its Firefly generative AI platform, including precision text-based video editing, camera motion controls, third-party upscaling via Topaz Astra, and the public beta launch of a new browser-based Firefly video editor.

Firefly Video Editor Public Beta

Let's start with the public beta of the new Firefly video editor. This is a browser-based editing environment designed to assemble generative clips, uploaded footage, music, and graphics into finished videos.

The editor supports a lightweight multi-track timeline as well as text-based editing for transcript-driven content such as interviews or talking-head videos. Export options range from vertical social formats to widescreen outputs intended to sit alongside traditionally edited material.

Rather than replacing the newly rebadged Premiere, Firefly video editor looks to be positioned more as an upstream creative assembly tool and somewhere where users can shape generative assets before moving into more complex post-production workflows.

Precision Video Edits Without Regenerating

Prompt to Edit for video allows creators to make targeted changes to an existing AI-generated clip instead of regenerating it from scratch.

Using text instructions powered by Runway’s Aleph model, users can issue commands such as removing an object or person from a frame, replacing a background, adjusting lighting conditions, or subtly reframing a shot. The key point here is that Firefly applies these changes directly to the existing clip, preserving elements that already work rather than rolling the dice on a new generation. As anyone that has ever tried to create imagery with AI knows, that is too often the surefire way to throw the baby out with the bathwater

Camera Motion Reference 

Firefly’s Video Model also gains a camera motion reference feature. Users can now upload a starting frame along with a reference video that defines the desired camera movement.

Instead of relying on text prompts alone to describe motion, Firefly uses the reference clip to anchor movement in a more predictable way. The aim is to deliver consistent, cinematic camera moves without repeated trial-and-error generations.

Video Upscaling via Topaz Astra

Firefly Boards now integrates Topaz Astra for video upscaling, allowing users to enhance footage up to 1080p or 4K.

The tool can be used to improve low-resolution clips for modern delivery platforms, restore older or archival footage, or bring legacy brand assets up to current quality expectations. Upscaling runs in the background, allowing users to continue working or queue additional clips without interrupting their workflow.

FLUX.2 Added

Adobe has also added FLUX.2 from Black Forest Labs to Firefly. The model supports photorealistic image generation, advanced text rendering, and up to four reference images.

FLUX.2 is available across Firefly’s Text to Image module, Prompt to Edit, Firefly Boards, and Photoshop desktop’s Generative Fill. Adobe says Adobe Express support will follow in January.

Unlimited Generations Until Mid-January

To encourage experimentation with all these new AI features , Adobe is offering unlimited image and video generations in the Firefly app until January 15 for users on Firefly Pro, Firefly Premium, 7000-credit, and 50,000-credit plans.

The offer covers Firefly’s own commercially safe image and video models, as well as partner image models including FLUX.2, Google Nano Banana, and OpenAI’s GPT Image. After January 15, standard credit limits will apply again.