Adobe adds new AI video tools and partner models to Firefly

Adobe has launched a set of new tools and model integrations for its Firefly application, expanding generative AI features for video creators.

The latest update introduces text- and voice-driven sound effect generation, new frame-by-frame video editing controls and integrations with partner AI models including Google’s Veo 3, Moonvalley’s Marey and Runway’s Gen-4.

The Firefly app’s new Generate Sound Effects feature, released in beta, allows users to create custom sound effects using either a written prompt or recorded voice. The tool supports impact and atmospheric sounds and enables layering with ambient audio or user-uploaded video.

Adobe also added frame-level video editing features for creators working in Firefly.

These include Composition Reference, which guides framing based on reference visuals; Keyframe Cropping, for shot refinement across sequences; and Style Presets, which apply consistent aesthetics to generated outputs.

The company has expanded its partner model offerings within the Firefly ecosystem. New models from Pika, Luma AI and Topaz Labs join those already available from OpenAI, Ideogram, and Black Forest Labs. All models can be accessed with a single Adobe sign-in and content created in Adobe apps will not be used for model training, per company policy.

As part of Adobe’s Content Authenticity Initiative, all AI-generated content in Firefly is tagged with Content Credentials, which disclose the model used to create the asset.

Adobe stated that Firefly models have been used to generate over 26 billion assets globally since launch.

Advertisement