Seedance 2.0 Prompts

A good video prompt should not leave you guessing what it might create. This Seedance 2.0 prompt collection lets you explore the finished videos first, so you can quickly find ideas with the motion, camera work, pacing, or visual style you want.

Browse trending examples discovered across public creator communities, from cinematic storytelling, realistic character performances, anime, advertisements, and product videos to surreal transformations, visual effects, and social media clips. Fresh examples are added regularly to reflect the styles and techniques creators are currently experimenting with.

Open any video to reveal its full prompt, view related examples, or copy the original wording. When you find an idea worth trying, select Use Idea to load the prompt into GPTProto. You can then replace the character, action, location, camera movement, atmosphere, or reference material and generate a Seedance 2.0 video of your own.

1. Focus on one main action. Short clips work best when the subject performs one clear action or transformation. Avoid combining multiple locations, storylines, or competing events in a single generation.

2. Describe events in chronological order. Explain what happens first, next, and last so Seedance 2.0 can follow the intended sequence instead of blending several actions together.

3. Separate subject motion from camera motion. State what the character or object does, then describe how the camera moves. For example: “The woman turns toward the window while the camera slowly tracks forward from behind.”

4. Specify the pace and ending. Words such as slowly, suddenly, continuously, or in one fluid motion help control timing. Describe the final pose, frame, or camera position when the ending matters.

5. Protect important visual details. When using a reference image, identify the face, clothing, product shape, background, or composition that must remain consistent throughout the video.

6. Match the camera style to the scene. Choose deliberate directions such as a close-up, low-angle tracking shot, handheld movement, slow dolly-in, or locked camera instead of asking for generic “cinematic motion.”

7. Review motion before adding more detail. If the first result has weak movement, refine the action, timing, and camera instructions first. Adding more style words will not necessarily correct an unclear motion sequence.