Adobe releases Firefly app for iOS and Android
Adobe has released a Firefly app for iOS and Android. It creates photos and videos using Adobe technology and partner models from OpenAI and others.
Until now, Adobe Firefly was only available as a web app. Now it also runs as a mobile app on smartphones.
(Image: Adobe)
Adobe has released a Firefly app, which will be available immediately for Android and iOS, as well as a web app. Firefly creates photos and videos, among other things, after prompt input. The service uses models developed by Adobe, but now also uses technology from third-party providers. Some partners will be added with the update.
Videos by heise
Integration of partner models
The Firefly service has been integrating image generators from third-party providers OpenAI, Google, and Black Forest Labs since April of this year. Adobe has now added AI models from Ideogram, Luma AI, Pika and Runway, which are primarily intended to help with video generation.
Under the Firefly umbrella, Google's text-to-image models Imagen 3 and 4, Flux 1.1 Pro and Flux.1 Context from Black Forest Lab, Ideogram 3.0, an OpenAI image generator and Runway's image model Gen-4 are available as image-generating AI.
In addition to the Firefly video model, Google's Veo 2 and 3 video generators are also available, which accept both text and images as input. These are joined by Ray2 from Luma AI and the text-to-video generator from Pika.
Functions of the Firefly app
The mobile app for Android and iOS generates images and videos after input from text prompts, in the case of the latter also from image prompts. It also includes the Photoshop tools Generative Fill for AI retouching and Generative Extend to extend the image area.
Adobe synchronizes all content created in the Firefly app with the user's Creative Cloud account.
Brainstorming with Firefly Boards
Adobe recently introduced Firefly Boards in public beta as an interface for brainstorming with generative AI. In addition to the image generator, Board now also supports video AI.
It should be possible to edit images iteratively on the board. To this end, Adobe is integrating the AI image processing Flux.1 context from Black Forest Labs and image generation from OpenAI.
Adobe documents can also be linked on the board. It synchronizes changes automatically. A clean-up function arranges all visual elements in a presentable layout.
Labeling AI-generated content
Adobe has integrated Firefly into its Content Authenticity Initiative (CAI). The service automatically provides generated content within the Firefly app with metadata, so-called “content credentials”, which identify it as AI so that the user knows whether it was created with a Firefly model or a partner model.
Adobe has trained Firefly with data from its agency service Adobe Stock, as well as with public domain content and content for which the copyright has already expired. This should minimize the risk of copyright infringements.
Prices and availability
Adobe charges for the use of Firefly via a credit model that is difficult to understand. Adobe Creative Cloud subscribers receive 1000 credits per month to generate images. For subscriptions to individual applications, there are 25 credits (InCopy, Substance 3D, Acrobat Pro), 100 credits (Lightroom), 250 credits (Express Premium) or 500 credits (Illustrator, InDesign, Photoshop, Premiere Pro, After Effects), depending on the type.
Users who only want to use the app, web app or video model can subscribe to Adobe Firefly from €10.98 per month. Firefly subscribers receive credits to use the video generator and unlimited access to the image generator.
(akr)
