Soutine Seedance Prompts
Prompt library: https://soutine.ai/seedance-prompts
Soutine's Seedance library pairs video prompts with example clips so you can study the shot before adapting the instructions. It focuses on describing subjects, camera movement, action, lighting, and a sequence of events rather than only the appearance of a still frame.
Open the Seedance landing page to preview a clip, copy its full wording, or load the prompt into the Soutine video generator.
What the library contains
At the time of this update, the page displayed 66 prompts with result previews. The collection can change. Examples include fantasy action, stylized urban scenes, product-like close-ups, fashion movement, and visual transformations.
The library supports browsing by frame ratio, including 16:9 and 9:16. Prompt cards offer a playable preview, Copy, and Use it. Additional cards can be revealed with Load more prompts.
The page's generator shares Soutine's workspace credits and task history. The landing page is a place to find and prepare a video idea, rather than a separate free inference service.
Intended use
| Goal | How the library can help |
|---|---|
| Short-form visual concepts | Study vertical framing and a single clear action. |
| Cinematic scene exploration | Inspect camera language, lighting, and wide compositions. |
| Product or campaign drafts | Start from a close-up or motion concept and replace the example subject. |
| Storyboarding | Examine how a prompt describes a beginning, a change, and an ending. |
| Learning video prompting | Compare an instruction with its preview and identify the choices that define the shot. |
These are suggested uses of the prompt resource. Each generated clip still needs review before inclusion in a finished project.
From a prompt card to a video
- Find a shot direction. Browse the cards and preview a clip. Choose a frame ratio that suits your intended use.
- Read the full instruction. Identify the subject, environment, camera move, action, and constraints.
- Copy or choose Use it. Copy provides the prompt text. Use it loads the prompt into the page's generator and selects a matching Seedance model; it does not itself submit generation.
- Rewrite for your idea. Change the subject and scene without losing the sequence of actions. Keep the requested camera behavior coherent.
- Review model and settings. Confirm the ratio and the available duration and resolution settings, then check the displayed credit quote.
- Generate and inspect. Follow the task in the workspace and watch the entire output, including transitions and the final frame.
Writing a useful video instruction
A video prompt benefits from a time-based description. Specify what appears at the start, what changes, and how the shot ends. State whether the camera is fixed or moving, and avoid asking for incompatible camera actions at once.
Illustrative prompt
This original example demonstrates a simple shot structure. It is not a tested library prompt, and no resulting clip is claimed:
A wide shot of a small paper boat on a quiet pond at sunrise. The camera stays fixed. A gentle ripple moves the boat toward the center, a leaf passes behind it, and the water settles. Soft golden light and restrained motion. No on-screen text.
The scene has one subject, one camera choice, and a small sequence of visible events. A constraint such as “no on-screen text” expresses a desired outcome but cannot guarantee compliance.
Tips for adapting a preview
- Choose the frame ratio before rewriting the composition. A vertical shot and a wide shot may need different subject placement.
- Keep one clear action or transformation as the main event.
- Make camera movement explicit: for example, a fixed shot, a slow push-in, or an orbit.
- Describe an ending so the instruction has a timeline rather than only a mood.
- Check the selected model after using a card; settings and model availability can change.
Limitations
The preview shows an example generation, not a guaranteed repeatable result. Motion, timing, physical interactions, and camera interpretation can vary. Subject appearance may drift across frames, and transitions or fine details may contain artifacts.
Changing models, aspect ratio, duration, or wording can alter the result substantially. A prompt pasted into another video tool may need adaptation; the Seedance preview is not a benchmark for that tool.
Video generation uses credits and can take time depending on settings and queue conditions. Review the current quote and task state before generating again. Check permissions for input media and the service's current output-use terms before publication.
Frequently asked questions
Can I browse without generating?
Yes. The page offers prompt text and clip previews to browse and copy. Generating a new clip is a separate workspace action.
What does Use it do?
It prepares the generator with the card's prompt and a matching model. You review the settings and confirm generation yourself.
Which frame ratio should I use?
Choose the ratio for your destination and composition. Use the library's filters to find previews already designed for that frame.
Will the result exactly match the preview?
No. The preview is a useful reference, but a new generation can differ in motion, appearance, and timing.
Does this repository provide Seedance weights or an API?
No. It is a README-only reference card; the live library and generation workflow are on Soutine.
Start with a clip you can inspect
Explore Seedance video prompts, preview a shot, and adapt its wording before submitting your own task.
Links
About this repository
The prompt library lives on the landing page. This is a README-only reference card for discovery and deep linking. It contains no model weights, training code, or hosted inference endpoint.