16:916:9Seedance 2.0 prompt library
5,926 prompts, spec-extracted and filterable.
Camera movement
Seedance 2.0 camera movement prompts →
Browse by category
cinematic realistic sports cinematic AI character fantasy short film vlog commercial cyberpunk anime 3D cartoon surreal animal nature food product urban music video abstract vehicle UGC vintage retro documentary game cinematic architecture interior tutorial stop motion chinese ink channel intro storyboard travel
About Seedance 2.0
Seedance 2.0 is a text-to-video and image-to-video model from ByteDance. It generates a short clip from a written prompt or a reference image, handles camera movement, lighting and stylisation, and can keep a subject consistent across shots.
Seedance 2.0 specifications
| Max clip length | 15s |
|---|---|
| Duration options | 5s default; up to 15s; -1 lets the model choose |
| Resolutions | 480p, 720p, 1080p, 4k |
| Aspect ratios | 16:9, 4:3, 1:1, 3:4, 9:16, 21:9, 9:21, adaptive |
| Max prompt length | 4000 characters |
| Native audio | yes — dialogue, sound effects and music |
| Reference inputs | up to 9 images, 3 videos (15s each), 3 audio files |
Verified against the vendor source on 2026-09-11 (source). Vendors change these frequently — confirm before relying on them. Where we could not verify a figure we say so rather than guess.
What 5,926 measured prompts show
Rather than describe the model in the abstract, here is what the prompts in this library actually contain. Camera language appears in 89.5% of them, lighting in 83%, and a consistency lock in 93% — the instruction that keeps a subject from drifting between frames.
The clearest pattern is duration: Seedance 2.0 accepts clips up to 15 seconds, and 71% of the prompts that state a duration ask for exactly 15 seconds. Creators write to the model's ceiling rather than to a spread of lengths. If you move a prompt to a model with a different ceiling, duration is the first field to revisit.
Computed from our own structured corpus (5,926 prompts, 1,196 authors), generated 2026-09-11. Full breakdown on the statistics page.
Frequently asked questions
What is a Seedance 2.0 prompt?
A Seedance 2.0 prompt is a written instruction describing the shot you want the model to generate: subject, action, camera movement, lighting and duration. Of the 5,926 prompts analysed here, 89.5% state camera language and 83% state lighting.
How long are typical Seedance 2.0 prompts?
71% of the prompts that state a duration ask for 15 seconds, making it the de facto standard clip length. 10 and 30 seconds are the next most common.
Which camera movements are used most often?
Tracking (2,443 prompts), close-up (2,347) and wide shot (1,547) are the most frequent. Prompt authors specify camera language far more often than aspect ratio.
Do these prompts work with other video models?
The wording is written for Seedance 2.0, but the shot structure — subject, movement, lighting, duration — transfers to other text-to-video models. Expect to adjust duration and camera complexity for models with shorter maximum clip lengths.
camera
lighting
duration
aspect
16:9
