Adobe is entering the world of generative AI video. Today, they’re launching the Firefly Video Model, which they’ve been talking about since earlier this year. This new model will be available in several new tools, including features in Premiere Pro that let creators extend video clips and create new videos from still images and text.
The first tool, called Generative Extend, is now available in beta for Premiere Pro. It helps users extend the beginning or end of clips that are a bit too short, or make changes during a shot, like fixing eye-line shifts or unexpected movements.
Clips can only be extended by two seconds, so Generative Extend is best for making small adjustments. This could save time by avoiding the need to re-shoot footage for minor problems. Extended clips can be created at either 720p or 1080p at 24 FPS.
The tool can also be used on audio to help make edits smoother, but there are some limits. For instance, it can extend sound effects and background noise by up to ten seconds, but it won’t work for spoken dialogue or music.

Adobe Launches New Text-to-Video and Image-to-Video Tools
Two new video generation tools are launching on the web. Adobe’s Text-to-Video and Image-to-Video tools, which were announced in September, are now available in a limited public beta in the Firefly web app.
Text-to-Video works like other video generators, such as Runway and OpenAI’s Sora. Users simply enter a text description of what they want to create. It can mimic various styles, including regular films, 3D animations, and stop motion. The generated clips can also be refined using different “camera controls” that simulate camera angles, motion, and shooting distance.

Image-to-Video takes it a step further by allowing users to add a reference image along with a text prompt for better control over the results. Adobe suggests this feature could be useful for creating b-roll from images and photos, or for visualizing reshoots by uploading a still from an existing video.
However, the example below shows that it can’t fully replace reshoots, as there are noticeable errors like wobbly cables and shifting backgrounds in the results.
You won’t be able to create full movies with this technology anytime soon. The maximum length for Text-to-Video and Image-to-Video clips is currently five seconds, and the highest quality is 720p at 24 frames per second.
In contrast, OpenAI claims that Sora can generate videos up to a minute long while keeping good visual quality and sticking to the user’s prompt. However, Sora is not available to the public yet, even though it was announced months before Adobe’s tools.
New AI Video Tools with Safety Features
Text-to-Video, Image-to-Video, and Generative Extend all take about 90 seconds to generate, but Adobe is working on a “turbo mode” to speed this up. Despite their limitations, Adobe claims that its tools, powered by its AI video model, are “commercially safe” because they’re trained on content the company has permission to use.
This is important since other providers, like Runway, are facing scrutiny for allegedly training their models on thousands of scraped YouTube videos or in Meta’s case, possibly even personal videos. For some users, the commercial viability of Adobe’s tools could be a key factor in their decision.
Another advantage is that videos created or edited with Adobe’s Firefly video model can include Content Credentials, which help disclose AI usage and ownership rights when published online. It’s uncertain when these tools will be fully released, but at least they are currently available to the public, unlike OpenAI’s Sora, Meta’s Movie Gen, and Google’s Veo generators.
The launch of these AI video tools was announced today at Adobe’s MAX conference, where the company is also unveiling several other AI-powered features for its creative applications.
Conclusion
Adobe’s new generative AI video tools offer exciting options for creators. With features like Generative Extend, Text-to-Video, and Image-to-Video, users can easily improve their video projects. Although there are some limitations, these tools provide new ways to create content and lead to future improvements in video technology.



