Seedance Prompt Generator
Describe the shot, pick a camera move and a style, and get a clean Seedance-ready prompt you can paste into Jimeng, Dreamina, or an API request.
Start from a preset
Thirty curated starting points. Pick one, then tweak the fields above. Each links to the article with more prompts in that style.
Builds prompts for use in Jimeng, Dreamina, or any Seedance API provider. This tool does not generate video.
How to use the Seedance prompt generator
A good Seedance prompt reads like a shot description, not a keyword list. ByteDance’s own prompt guidance for Seedance 2.0 recommends describing the subject, the action, the scene, and a single camera movement per shot, and writing multi-shot sequences as a numbered storyboard (“Shot 1 / Shot 2 / Shot 3”). This Seedance prompt generator bakes that structure in: fill in the fields, pick presets for camera, style and lighting, and it assembles one coherent paragraph you can paste into Jimeng, Dreamina, Doubao, CapCut, or an API call to Volcano Engine, BytePlus, fal, Replicate, or Together AI.
- Subject and action. Be concrete. “A lone astronaut in a weathered orange suit walks slowly across a red desert” gives the model far more to work with than “astronaut on Mars”.
- Scene. Add time of day, location and atmosphere. Lighting and style presets layer on top of this.
- One camera move. The official guidance warns against stacking camera moves in a single shot. If you want several, switch to a 2–4 shot sequence and the tool assigns a move per shot.
- Audio. Seedance 1.5 Pro and every 2.x model generate sound with the picture. Describe dialogue in quotes, plus ambience and effects; leave the toggle off for a silent clip and the prompt will say so.
- Avoid list. Seedance has no dedicated negative-prompt field in most apps, so the generator appends an “Avoid:” sentence instead.
Duration, aspect ratio and references are settings, not prompt text
The generator shows a settings line under every prompt because clip length and aspect ratio are chosen in the app or API, not inside the prompt. On the BytePlus ModelArk API the Seedance 2.0 series accepts 4–15 seconds and Seedance 2.5 accepts 4–30 seconds, both at 24 fps, with aspect ratios of 21:9, 16:9, 4:3, 1:1, 3:4 and 9:16. Reference images, video and audio clips are uploaded separately and then referred to in the prompt with tags such as “@Image 1” in Dreamina or “[Image 1]” in API samples; see our image-to-video prompt guide for that workflow.
Presets and where to go next
The thirty presets cover product ads, cinematic shots, anime, realistic characters with dialogue, image-to-video, multi-shot stories, real estate, fashion, travel and music video. Each one links to the article with more prompts in that style, starting with our main Seedance prompts library. If you are deciding which version to use, the Seedance 2.5 guide and the Seedance 2.0 guide explain the differences, and the cost calculator estimates what your prompts will cost at scale.
Frequently asked questions
Does this tool generate video?
No. It assembles a text prompt in a structure that works well with Seedance and runs entirely in your browser. You paste the result into Jimeng, Dreamina, Doubao, CapCut, or an API request to a provider such as BytePlus ModelArk, fal, Replicate or Together AI, and that service generates the video.
Is the prompt generator free?
Yes. There is no account, no limit, and nothing you type is sent to our servers. Seed Video is an independent publication funded by display advertising and has no affiliation with ByteDance or any provider.
Which Seedance versions do the prompts work with?
The structure works across Seedance 1.x, 2.0, 2.0 Fast, 2.0 Mini and 2.5. Audio descriptions only apply to models that generate sound: Seedance 1.5 Pro and the 2.x series. Duration limits differ by version, so check the settings line under the output before you generate.
How do I reference an uploaded image or video in the prompt?
Upload the asset in the app or API first, then mention it with the platform’s tag. Dreamina uses labels like @Image 1 and @Video 1; BytePlus code samples use [Image 1] and [Video 1]. The 2.0 series accepts up to 9 images, 3 videos and 3 audio clips; Seedance 2.5 raises that to 30 images, 10 videos and 10 audio clips.
Sources: BytePlus ModelArk Seedance 2.0 prompt guide and Seedance 2.5 tutorial (retrieved August 23, 2026). Seed Video is not affiliated with ByteDance; product names are used for identification only.