Adobe's magical AI video creation tool is now available

Now available on a redesigned web app, Firefly is a new product in Adobe's AI services line that aims to let users create videos that are commercially safe, meaning they can be used for professional purposes.

Adobe Firefly can turn still photos into videos using AI.

Firefly allows users to create 5-second AI videos in 1080p resolution, but without sound. Users can enter text prompts or use images to create footage. Firefly's interface comes with a control panel that allows you to customize the aspect ratio, footage size, camera angle, frames per second, and motion effects like zooming in, zooming out, or tilting.

Adobe says its AI models are trained on content from Adobe Stock and other public sources, not using user work. Adobe hopes that AI video will help create new clips from existing footage, while also helping to save production costs for aerial footage instead of relying on drones.

The beta version of Premiere Pro also allows users to export a frame into Firefly to create a video from a reference image, which is useful for those working on a specific project. Additionally, video editors can extend content by 2 seconds using the extender tool in Premiere Pro.

Firefly is Adobe's answer to OpenAI's Sora.

Last December, OpenAI launched its AI video generator, Sora, for paying ChatGPT users. Other AI video technologies like Meta's Movie Gen and Google's Veo have also been showcased at tech events and are expected to be publicly released this year. However, direct comparisons between these models are not yet possible.

One notable difference between Sora and Firefly is that Adobe's AI video model doesn't have a watermark indicating it's an AI product, although AI content in Premiere Pro is still marked on the timeline.

Users can experience the new AI visualization tools through two Firefly plans: the $10/month Standard plan offers 2,000 images and up to 20 AI videos, while the $30/month Pro plan offers 7,000 images and up to 70 AI videos. Both plans allow for unlimited AI visualizations.