On the Adobe Firefly homepage, select Video > Generate video.
Last updated on
30 Jun 2026
Learn to use Seedance 2.0 and Seedance 2.0 Fast to generate videos in Adobe Firefly.
Seedance 2.0 and Seedance 2.0 Fast are partner models available in Adobe Firefly that generate high-quality, reference-driven cinematic videos from text prompts and reference media. As omni models, they support both Text to video and Image to video workflows and can use up to 12 reference assets to help guide the generated video.
You can use reference assets to provide creative direction for visual style, characters, motion, composition, and audio, that allow greater control and consistency across generated videos.
Start video generation
In the General settings section, select Seedance 2.0 or Seedance 2.0 Fast from the Model dropdown menu.
Note
- When you select Seedance models for video generation, your prompts and reference files are sent to the model provider for processing. Adobe does not use your content to train generative AI models. The first time you use the Seedance models, review the partner model information prompt and select OK to continue.
- When using reference media with the Seedance models, uploaded images and videos cannot contain real human faces. If a real human face is detected in a reference image or video, Firefly displays a notification and prevents the asset from being used as a reference. However, you can still generate videos that include human faces by describing them in a text prompt or use a different model.
Select Resolution and Aspect ratio as per your preference.
When working with Seedance 2.0 Fast, you can quickly generate a lower-resolution (480p) video and then use Seedance 2.0 to generate a higher quality video.
From the Duration dropdown select a video length between 4 and 15 seconds.
If you want synchronized audio, enable the Audio toggle.
In the References section, select Frames > First to upload a starting frame that guides the opening scene and Frames > Last to upload an image as the ending frame of the generated video. To further modify a keyframe image, hover over it and select:
- Crop Image: If the keyframe image doesn’t match the aspect ratio, hover over the thumbnail to activate the image tools and select the Crop Image icon . You can move your cursor over the preview and adjust the crop as needed.
- Edit Image: If you wish to edit the first or last frame image, hover over the thumbnail to activate the image tools and select the Edit Image icon . When the edit with text prompts page opens, you can easily add or modify the details of your image and Save it to get back to the Generate image page.
- Remove Image: To remove an image from the first or last frame, hover over it and select the Remove Image icon .
In the References section, select Media and upload up to 12 reference assets — including up to nine images, three videos, and three audio files — from Your device or Adobe cloud storage.
When you upload an asset, it gets a default name based on its media type. For example, the first image you upload becomes Image1, the first video becomes Video1, and the first audio file becomes Audio1. You can reference any uploaded asset in the prompt bar using an @-mention.
Type @ followed by the asset name to call it up. This lets you combine multiple assets in a single prompt without re-uploading or describing them. For example - "@Image1 is a person holding the object in @Image2."
If you want to generate similar results across multiple generations, enter a Seed value in the Advanced settings section.
In the Prompt field, enter a detailed description of the video you want to generate.
Select Generate.
Turn text into art with Adobe Firefly
Create beautiful images and video clips from text prompts using generative AI.
