seedance-prompt-zh

Generate multi-modal Seedance 2.0 prompts using @-reference syntax.

Updated May 25, 2026

Search Tags:#prompt engineering #multimodal #creative workflow #video generation #ai video #seedance

Ready to run

npx skills add https://github.com/fuyucn/skills --skill seedance-prompt-zh-fuyucn

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: seedance-prompt-zh
Source: https://github.com/fuyucn/skills/tree/main/seedance2
Command: npx skills add https://github.com/fuyucn/skills --skill seedance-prompt-zh-fuyucn

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill solves the complexity of prompting multi-modal AI video models by providing a structured framework for Seedance 2.0, ensuring high-quality, coherent, and technically accurate video outputs.

Core Features & Use Cases

Multi-modal Prompting: Expert guidance on using @-reference syntax for images, videos, and audio.
Advanced Control: Precise instructions for camera movement, character consistency, and scene transitions.
Use Case: Create a 15-second cinematic product advertisement by referencing a product image, a specific camera movement video, and a background music track to ensure perfect synchronization.

Quick Start

Use the seedance-prompt-zh skill to generate a prompt for a 10-second cinematic video featuring a character from an uploaded image performing a specific action referenced in a video file.

Frequently Asked Questions about seedance-prompt-zh

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I write multi-modal prompts for AI video generation with image and audio references?▼

Multi-modal AI video generation prompts use specific @-reference syntax to integrate images, videos, and audio. This ensures precise audio-visual synchronization and character consistency by directly linking external assets within the text instructions.

What is the best way to maintain character consistency across different video generation scenes?▼

Character consistency in video generation is achieved by referencing a specific source image using the designated @-reference syntax. This anchors the subject's visual identity, ensuring the AI model replicates the exact features across multiple generated scenes.

Can I replicate exact camera movements from a reference video in my AI generated video?▼

Yes, you can replicate camera movements by using a reference video file within your prompt. The multi-modal framework extracts the motion trajectories from the referenced video and applies them to your newly generated scene.

How do I create a cinematic product advertisement using multi-modal video generation?▼

To create a cinematic product advertisement, you reference a product image, a camera movement video, and a background music track simultaneously. This multi-modal workflow ensures perfect synchronization between the product visuals, motion, and audio.

Does the Seedance 2.0 video model support audio-visual synchronization in prompt generation?▼

Yes, audio-visual synchronization is supported by referencing audio tracks alongside visual assets in the prompt. The framework aligns the audio cues with the generated video timeline to ensure perfectly matched multi-modal outputs.

What are the input constraints when generating multi-modal video prompts for complex workflows?▼

Generating multi-modal video prompts requires strict adherence to specific input constraints and @-reference syntax for asset integration. You must properly format image, video, and audio references to ensure the AI model processes the complex workflow correctly.