How to prompt Gemini 3.1 for Epic animations

Gemini 3.1 Pro for Animation: A Detailed Process for Production-Ready Results

Key Concepts:

• Scene-Based Prompting: Structuring animation requests around individual scenes with detailed timing, UI states, and actions.
• Planning vs. Building: Separating the animation process into distinct planning and execution phases.
• UI Cloning: Replicating existing UI components for use in animation creation.
• Skill Libraries: Pre-built prompting structures designed for specific animation tasks (e.g., product launch demos).
• Gemini 3.1 Pro: Google’s latest large language model, noted for improved animation capabilities.

1. The Problem with Current Gemini 3.1 Pro Animation Results

The speaker observes that while Gemini 3.1 Pro is touted for its animation capabilities, many users are achieving results resembling basic PowerPoint slides – far from production quality. This discrepancy arises from how prompts are constructed. Vague, open-ended requests like “animate this” or “make a prop” lead to chaotic and uncoordinated animations because the model struggles with the complex spatial reasoning required for effective animation.

2. The Core Solution: Structured Scene-Based Prompting

The key to improving animation quality lies in separating the planning and building stages. The speaker advocates for a scene-based prompt approach. This involves meticulously outlining each scene of the animation, detailing the timing, UI states, and desired actions. This structured approach removes the burden of spatial planning from the model, allowing it to focus on implementation. The speaker guarantees a “two to three-x better result” simply by adopting this method, even without advanced prompting techniques.

3. Components of a Robust Animation Prompt

A well-crafted animation prompt, according to the speaker, consists of four key elements:

• Timing: Specifying the duration of each scene or action.
• Clearly Defined State: Describing the exact appearance and configuration of elements within each scene.
• Special Keyword Effects: Incorporating terms to instruct specific animation behaviors (e.g., “3D perspective,” “rotation,” “fading with sector delay,” “tagline with standard item reviews”).
• Action: Describing what is happening in the scene.

4. Case Study: Rocket’s Open Claw Video

The speaker references a video created by Rocket showcasing Open Claw as an example of successful scene-based prompting. The video’s superior quality, compared to typical product launch animations, is attributed to the prompt’s structure. The prompt breaks down the animation into scenes, specifying timing, state details (what should be in the frame), and special effects keywords. This allows the model to follow instructions and produce a high-quality result.

5. Real-World Application: Product Launch Animation with Super Design

The speaker details their own experience creating a product launch animation video for Super Design’s “Inspire Mode” using the described methodology. They created scene-based prompts detailing timing, special effects, state, action, and technical details for each scene. This resulted in a polished product launch video.

6. The Super Design Skill Library & Workflow

To streamline the prompt creation process, the speaker developed a “skill” within the Super Design skill library specifically for product release demo animations. The workflow involves:

• UI Cloning: Utilizing Super Design’s free Chrome extension to clone UI components from websites, generating clean HTML code.
• Skill Activation: Selecting the cloned UI and activating the “Product Release Demo” skill within the Super Design library.
• Prompting & Refinement: The skill prompts the user for animation details (initial state, tab behavior, inspiration item pop-up, cursor appearance). It then generates a detailed animation plan with a prompt breakdown, including timing and UI state information.
• Building & Generation: Once the plan is approved, the skill generates the final animation video.
• Gemini 3.1 Pro Mode: The speaker recommends using Gemini 3.1 Pro mode within the skill for optimal animation quality.

7. Accessibility & Generalizability

The speaker emphasizes that the described prompting process is not exclusive to Super Design. It can be applied with other AI tools like Cloud Code or Cursor. Super Design’s skill library simply provides a more user-friendly and streamlined implementation of the core principles.

8. Data & Statistics (Implied)

While no specific numerical data is presented, the speaker’s claim of a “two to three-x better result” with scene-based prompting suggests a significant improvement in animation quality based on their experimentation. The success of the Super Design Inspire Mode launch video serves as a practical demonstration of the methodology’s effectiveness.

9. Notable Quote:

“One of the key reasons your animation generated by model is bad is because you're giving a vague and open-ended request like animate this or make a prop which leads to chaotic uncoordinated motions and efforts.” – The speaker, highlighting the importance of specific prompting.

10. Synthesis & Conclusion

The key takeaway is that achieving production-ready animations with Gemini 3.1 Pro requires a shift from vague requests to meticulously planned, scene-based prompts. By separating the planning and building phases, and by detailing timing, UI states, and desired effects, users can significantly improve the quality and coherence of their animations. Tools like Super Design’s skill library can further streamline this process, but the underlying principles are applicable across various AI platforms. The emphasis is on providing the model with clear, structured instructions to overcome its limitations in spatial reasoning and deliver compelling visual results.