higgsfield-generate

Image-to-video generation allows you to create UGC-style ad videos by passing a product image as a media input. The Skill uses Seedance 2.0 defaults to generate high-fidelity branded video content directly from your reference.

Yes, you can use a reference image to generate a new image via the image-to-image workflow. The Skill processes your media input alongside a text prompt to produce a new image with sensible production defaults like GPT Image 2.

The best way to create branded marketing content with avatars and hooks is using Marketing Studio constructs. The Skill guides you to produce branded ad content by defining avatars, products, hooks, and settings in a single generation flow.

Yes, Higgsfield supports audio-reference matching for video generation when using the Seedance 2.0 model. You can provide an audio reference for lipsync or soundtrack matching during the image-to-video workflow.

To get the final result URL without manual polling, the Skill uses a single-shot create-with-wait execution method. This submits your request via the Higgsfield CLI and waits until the final media URL is returned directly.