How to Use Seedance
Seedance turns a sentence — or a single image — into a short video clip. Follow these five steps to make your results look intentional.
- 1
Choose your starting point
Text-to-video builds a scene from a description. Image-to-video animates a still you upload. Pick based on whether you already have the first frame.
- 2
Name the subject
Start with what's in frame — a person, product, place, or creature. Be specific about the one thing the shot is about.
- 3
Give the camera a job
Add a movement: slow dolly in, aerial drone descending, 360° turntable, handheld follow. Motion is what separates video from a moving photo.
- 4
Set lighting, mood & style
Golden hour, dramatic rim light, soft studio key — then a finish like 'cinematic, anamorphic lens flare' or '2D hand-painted anime'.
- 5
Generate, then iterate
Run it, change one thing at a time, and re-run. You'll quickly learn what each edit does to the result.
How to use Seedance 2.0, step by step
Seedance 2.0 is the version you'll actually run today. The flow:
- Pick where to run it. Dreamina and CapCut are the consumer apps; ponpon runs Seedance with no setup or API keys; fal, Replicate, and BytePlus ModelArk serve the API for pipelines. Details in how to access Seedance 2.0.
- Write the prompt. Use the four-part structure below — subject, camera, lighting, style.
- Add references if you need control. 2.0 accepts multimodal references in one prompt — roughly 9 images, 3 videos, and 3 audio clips — to lock characters, style, or motion.
- Want speech? Write the line. 2.0 generates voice and lip-synced dialogue in the same pass — put the spoken line in quotes in your prompt. See the lip-sync guide.
- Pick an aspect ratio and generate. 16:9 for landscape, 9:16 for vertical, 1:1 for product loops — then iterate one change at a time.
Text-to-video vs image-to-video
- Text-to-video (t2v): you describe a scene from scratch. Best for original ideas, establishing shots, and stylized looks.
- Image-to-video (i2v): you upload a still and describe how it should move. Best for animating a specific character, product, or photo while preserving its look.
Pick based on whether you already have the “first frame” you want.
A prompt structure that works
Strong Seedance prompts stack four things:
- Subject— what's in frame (“a lone astronaut”, “a matte-black earbud case”).
- Camera & motion— give the camera a job (“slow dolly in”, “aerial drone descending”, “360° turntable”).
- Lighting & mood— “golden hour”, “dramatic rim light”, “soft studio key”.
- Style & finish— “cinematic, anamorphic lens flare, film grain” or “2D hand-painted anime”.
Example: “Cinematic aerial drone shot flying through a neon-lit city at night, rain-slicked streets, camera slowly descending, anamorphic lens flares, moody atmosphere.”
Going deeper? The Seedance prompt guide covers all five elements — including the audio layer — with basic-to-best examples.
Tips for better results
- Name the motion. Static descriptions produce static-feeling clips — add wind, particles, or a camera move.
- Keep i2v motion subtle (“gentle head turn”, “hair moving slightly”) and say “preserve identity” to avoid face drift.
- Iterate one variable at a time — camera, then lighting, then style — so you learn what each edit does.
- Match aspect ratio to the platform — 16:9 landscape, 9:16 vertical, 1:1 product loops.
Start from a proven prompt
View all →
Snappy Cartoony Anime Acting Style Prompt
A style prompt adding snappy, exaggerated pose-to-pose motion to anime characters.
by @suji_pop
Anime Soccer Player Tiger Spirit Goal
A female soccer player unleashes a giant tiger spirit to score an epic stadium goal.
by @Just_sharon7
Dark Fantasy Knight Cavalry Charge
A cinematic dark-fantasy storyboard of an armored visor-down knight charging into battle on horseback.
by @YaZoraiz
Frequently asked questions
What makes a good Seedance prompt?
A clear subject, an explicit camera movement, defined lighting/mood, and a style keyword. Stacking those four consistently produces usable clips.
What's the difference between text-to-video and image-to-video?
Text-to-video generates a scene from your description. Image-to-video animates a still image you provide, preserving its composition while adding motion.
Why does my video look static?
Add explicit motion — a camera move, wind, particles, or subject action. Descriptions without movement tend to produce static-feeling clips.
How do I keep a character consistent in image-to-video?
Describe only subtle motion (a gentle head turn, slight breeze) and add 'preserve identity' to your prompt. Start from a sharp, well-lit source image.
How do I use Seedance 2.0?
The same flow as any Seedance version: pick a platform (Dreamina, CapCut, ponpon, or the API), write a prompt with a clear subject, camera move, lighting, and style, then iterate. Seedance 2.0 adds multimodal references and native audio with lip sync — write dialogue into the prompt and it generates the voice too.
Can I use Seedance 2.0 for free?
Partly. Dreamina gives a small daily free credit allowance (outputs carry a watermark), and BytePlus has offered a free token grant for new API accounts. Free tiers are enough to practice prompts; regular or watermark-free generation needs a paid plan.