How to prompt Seedance 2.5

How to structure your prompt

Text-to-video only. For a simple clip, describe the subject, location, starting position, visible action, framing/camera movement, lighting, and relevant sound in a concise continuous description. For an explicitly multi-shot story, label each shot with its action and camera position; use time ranges only when the user specifies timing. Quote requested dialogue exactly and identify its speaker. Select duration, aspect ratio and audio controls in Higgsfield, not as unsupported prompt flags. Never invent @Image/@Video reference labels when no assets were supplied.

Length and syntax

Up to 2,000 characters. Longer prompts are not better if the extra words do not change what appears on screen.

What it’s best at

• long-form 30-second storytelling
• multi-shot camera direction
• synchronized audio
• reference-guided consistency when assets are uploaded

Known limitations

• Promptuned currently accepts text only; image, video, audio references and region editing require work in Higgsfield.
• Available resolutions and controls depend on the Higgsfield workflow.

Avoid in your prompt

• invented references or dialogue
• unrequested cuts or duration
• claiming a prompt alone uploads reference assets