muapi-seedance-2

What This Skill Does

muapi-seedance-2 is an advanced video generation skill designed for OpenClaw agents to interface with ByteDance’s Seedance 2.0 engine. It transforms simple user requests into professional 'Director Briefs' that utilize precise camera grammar, lighting physics, and temporal controls. Unlike standard generative models, this skill treats the agent as a Director, allowing for complex multi-modal orchestration including text-to-video, image-to-video (supporting up to 9 references), and seamless video extension. It manages character consistency, style locking via tagging, and high-fidelity synchronized audio generation.

Installation

To integrate this skill into your agent, run the following command in your terminal: clawhub install openclaw/skills/skills/anil-matcha/muapi-seedance-2

Use Cases

• Pre-visualization: Create cinematic storyboards and moving concepts for film and advertising.
• Content Automation: Rapidly iterate on B-roll or motion graphics using source imagery for brand consistency.
• Temporal Expansion: Extend existing clips while maintaining structural integrity and camera movement paths.
• Style Transfer: Apply the visual aesthetic of uploaded reference images to newly generated cinematic content.

Example Prompts

1. "Generate a 10-second cinematic shot of a neon-lit cyberpunk market. Use @image1 for the character's clothing and style, and ensure the camera performs a slow tracking shot through the crowd."
2. "Create a video of an ancient temple reveal. Start with a wide drone shot and finish with a close-up on the entrance. Ensure the mist effects are volumetric and the color grade is epic."
3. "Extend the last 5 seconds of @video1. Keep the same character in the scene, but make them walk through a doorway into a brighter room with sub-surface scattering lighting."

Tips & Limitations

• Order Matters: Always place your most critical composition instructions at the very beginning of the prompt.
• Technical Vocabulary: Avoid subjective terms like 'pretty' or 'good'. Instead, use terms like 'shallow depth of field', 'anamorphic lens', or 'volumetric lighting'.
• Reference Limitations: While you can use up to 9 images for i2v, balance is key; if images conflict, the model may struggle with identity stability. Use clear, distinct shots for best results.
• Duration: While the model supports flexible lengths, coherence is highest within the standard 5-10 second burst cycles.