How to Write Seedance 2.0 Prompts? NadouPro Prompt Library: Five Viral Video Formats and Full Prompts

Seedance 2.0 is one of the strongest models in the NadouPro video model matrix for multi-shot storytelling. This article collects five video formats that perform best on social media, each with a complete prompt that creators can copy, adjust, and use directly.

One rule applies to all the formats below: declare the shot structure at the beginning of the prompt. When using Seedance 2.0, write the number of shots, total duration, and aspect ratio at the top of the prompt, then build the rest of the content around that.

Format 1: Transformation

Transformation is currently the highest-performing format on Seedance. The key is to write shot by shot—number each shot, describe the precise action in it, and give Seedance a clear progression arc: calm → threat → transformation → aftermath.

Example: Mutant Girl

Take a typical scene as an example. A girl eating a hamburger mutates on the hood of a pickup truck into a giant monster, swallows the approaching zombie, then turns back into human form and continues eating the burger. A dark comedy rhythm with a horror undertone. 6 shots, 15 seconds. Requires 4 reference images (girl, truck, scene, zombie reference).

Montage, multi-shot action film, do not use a single camera position or a single cut, cinematic lighting, photorealistic, 35mm film texture, professional color grading, sharp focus, highly detailed textures, film grain, depth-of-field control, cinematic camera aesthetics. A pink-haired girl with glasses, wearing a cream top and jeans, sits on the hood of a white pickup truck beside a concrete overpass at dusk, eating a hamburger. A pale zombie bursts out from the shadows of the river channel. The girl puts down the hamburger, her body erupts into a giant pale fanged monster with extending limbs and clawed hands, swallows the zombie whole, then shrinks back into human form and picks up the hamburger. Handheld shake throughout, dark comedy rhythm with a horror undertone. Shot 1: medium shot of the girl sitting cross-legged on the hood, lazily chewing the hamburger, golden dusk light reflecting on her glasses and pink hair. Shot 2: wide-angle shot of the concrete river channel as the zombie bursts from the shadows under the bridge and sprints forward. Shot 3: close-up of the girl noticing the zombie, chewing slows, eyebrows lift slightly, she puts down the hamburger. Shot 4: medium shot of the girl jumping down from the hood, her body violently swelling and twisting into a giant pale fanged monster. Shot 5: wide-angle low-angle shot of the monster charging forward, grabbing the zombie with huge claws, lifting it, and swallowing it. Shot 6: medium shot of the monster shrinking back into the girl, jumping back onto the hood, picking up the hamburger, and continuing to chew. Total: 15 seconds / 6 shots / 16:9

Pro tip: For monsters and content that needs realism, if the skin looks too smooth or plastic, add "no 3D, no cartoon, no special effects" to the prompt. This forces the model into a hyper-realistic mode.

Format 2: POV Shot

A POV shot is a single continuous first-person shot—the camera is the character's eyes. The formula is always the same: one shot, first-person perspective, ultra-chaotic handheld movement, 15 seconds. What changes is the power, the environment, and the enemy.

Example: Lightning-Powered Character

Take a typical scene as an example. From a first-person perspective, a character with lightning powers fights in storm-soaked industrial ruins and ultimately destroys a giant steel titan by overloading its core. Single continuous shot, 15 seconds.

Pro tip: Use square brackets in the action description to specify effects—[Effect: branching circuits flicker with white-blue electricity]. This tells Seedance the specific manifestation of the power on the character's body without interrupting the shot description.

Format 3: Combat

The combat format emphasizes physical choreography and power contrast. Unlike the transformation format, characters in combat keep their original forms and create tension through moves, the environment, and rhythm.

Example: Train Fight

Take a typical scene as an example. A character fights an opponent on a moving train, using environmental elements such as the roof and the connections between carriages. The shot numbering is clear, and each action segment has a beginning, development, turn, and resolution.

Pro tip: For comedy scenes, add "add a visual gag in the background" and Seedance will create one on its own; you do not need to describe it.

Format 4: Animation

Seedance 2.0 is not limited to realistic styles; it is also strong at animated formats. From Japanese anime to French graphic novels, the key to different styles is to clearly define the visual language in the prompt.

Example: Desert Hero (Anime Style)

Take a typical scene as an example. A character adventures in a desert environment, with an overall bold-line anime illustration style—flat cel shading, thick solid confident outlines, warm amber and soft tones. In the prompt, state "the entire video must match this 2D anime illustration aesthetic, and all frames must remain consistent."

Example: Mixing 3D and 2D

Alternate between different animation styles within the same project. The key is to clearly declare the style transition at the beginning of each shot, so the model understands it is an intentional design rather than an error.

Format 5: Scene Continuation

To continue an existing clip, upload the previous video as context and write the next beat. The same face, the same clothing, and seamless continuity from shot to shot. This format does not require a new starting frame; it uses an @ video reference to continue the scene.

Prompt Writing Essentials

Shot numbering. In multi-shot sequences, use "Shot 1" and "Shot 2" to number them, and describe each shot separately. Seedance will execute them in numbered order.

Camera movement terms. Use cinematic camera language directly: dolly-in close shot, move left, arc camera movement, handheld tracking shot, rising crane shot. The model reads these terms directly.

Progression arc. The transformation format follows the "calm → threat → transformation → aftermath" arc. The combat format follows the "standoff → exchange → turning point → finish" arc. Reflect this rhythm in the prompt.

Reference citations. Use @ tags to reference uploaded character images, scene images, motion references, or audio. The model treats them as generation constraints.

Style lock. In animation formats, state the aesthetic style in the prompt and emphasize that "all frames remain consistent." In realistic formats, add "no 3D, no cartoon, no special effects" to force hyper-realism.

FAQ

Q: How many shots can a Seedance 2.0 multi-shot sequence support at most?

Seedance 2.0 supports multi-shot sequences up to 15 seconds long, and the number of shots depends on the duration of each shot. A typical configuration is 6 shots × 2.5 seconds, or 3 shots × 5 seconds.

Q: Should prompts be written in Chinese or English?

NadouPro supports Chinese prompts. Camera movement terms can remain in English (such as dolly in and tracking shot), and the system will recognize them correctly. It is recommended to write the main description in Chinese and keep technical terms in English.

Q: How can I keep characters consistent across multiple shots?

Use the @character tag to reference a character reference image. If you need to maintain consistency across multiple independent generations, use image-fusion mode to train an identity model.

Q: What is the difference between the transformation format and the combat format?

In the transformation format, the character changes form (from human to monster and back again), emphasizing a progression arc and visual impact. In the combat format, the character keeps the original form and creates tension through move choreography and environmental interaction. Both use multi-shot sequences and a numbered structure.

Q: How should prompts for an animation style be written?

State the visual style at the beginning of the prompt (such as "bold-line 2D anime illustration style, cel shading"), and emphasize at the end that "all frames keep this style consistent." When different shots need to switch styles, clearly state the style transition at the beginning of each shot.

Q: How should unsatisfactory generation results be handled?

Change only one variable at a time—the prompt or the reference image, not both at once. If the issue is skin texture, add "no 3D, no cartoon, no special effects." If the issue is camera movement, replace vague descriptions with more specific camera movement terms. If the character drifts, add an @character reference or use image-fusion mode.

For more features, visit NadouPro: NadouPro

Contains AI-generated information that may be incorrect. Please verify independently.