First-to-last transition
Example prompt
Move smoothly from the daylight first frame to the night-time last frame; keep the building geometry stable while lights turn on one floor at a time.
credits : 1
Seedance 2 AI Video Generator
Turn a written idea or reference image into an AI video with Seedance 2. Choose Standard or Fast, then use text-to-video, first-frame image-to-video, or first-and-last-frame generation.
How to Create a Video with Seedance 2
Create a Seedance 2 video in three straightforward steps using the generation modes and settings currently available in the tool.
- Choose Your Mode: Start with a text prompt, upload one image as the first frame, or upload two images to guide both the first and last frames.
- Describe the Motion: Add the subject, action, camera movement, lighting, pacing, and visual style you want. Then choose Seedance 2 Standard or Fast and review the available settings.
- Generate and Refine: Start the generation, review the result, and adjust the prompt or reference frames when you want a different motion, composition, or ending.
Seedance 2 AI Video Generator FAQs
What can I create with Seedance 2 on this page?
You can generate a video from text, animate a single image as the first frame, or use two images to guide the first and last frames. The generator currently offers Seedance 2 Standard and Seedance 2 Fast.
What is the difference between Seedance 2 Standard and Fast?
Both options support the Seedance 2 workflows exposed in the generator. Fast is available for quicker iteration, while Standard is the alternative for a regular generation workflow. Review the current settings and credit estimate before generating.
Can I use Seedance 2 for text-to-video?
Yes. Choose text-to-video, enter a prompt describing the scene and motion, select Standard or Fast, and use the settings currently shown in the generator.
How does Seedance 2 image-to-video work?
Upload one image to use it as the first frame, then describe how the scene should move. You can also upload a second image as the last frame to guide how the video begins and ends.
Do I need two images for first-and-last-frame video generation?
Yes. The first image guides the opening frame and the second guides the ending frame. If you only upload one image, the generator uses it as the first-frame reference.
What settings are available for Seedance 2?
The available duration, resolution, and aspect-ratio options are shown directly in the generator and may differ by mode. Choose from the settings currently displayed before starting a task.
How should I write a Seedance 2 prompt?
Describe the subject, action, environment, camera movement, lighting, mood, and pacing. For image-to-video, focus on the motion you want while keeping the reference-frame composition in mind.
Does this Seedance 2 page support audio or video-reference input?
The current page focuses on text-to-video and first-frame or first-and-last-frame image-to-video. Audio generation and video-reference input are not exposed in this workflow.
Current settings and limits
Seedance 2 Standard and Fast workflows for text, first-frame, and first-and-last-frame video generation.
- Input
- Text, one first-frame image, or first and last frames
- Models
- Seedance 2 Standard and Fast
- Duration
- Standard: 5/10/15 seconds; Fast: 4/8/15 seconds
- Resolution
- 480p or 720p
- Prompt limit
- Up to 2,500 characters
Model result Gallery
The Inspire panel above shows only public results saved for models in this Hub.
Model-specific prompt examples
Character action
Example prompt
A cyclist stops beside a mountain overlook, removes their helmet, and looks toward the valley; slow tracking camera, natural wind, restrained motion.
Product sequence
Example prompt
A glass perfume bottle emerges from shallow water as the camera arcs from left to right; crisp reflections, soft blue lighting, elegant pacing.
Know before you generate
Mode-specific settings
Duration choices differ between Standard and Fast, so review the controls after switching models.
Reference-frame drift
First and last frames guide the sequence but do not guarantee exact identity, geometry, or transition details.
Focused workflow
Audio generation and video-reference input are not exposed on this page.