KinoViz Team
One of KinoViz's most powerful features is AI Animation Generation — describe what you want in plain English, and get a complete animation project in seconds.
How It Works
KinoViz's AI generation is powered by Google Gemini 3.1 Pro Preview, a large language model optimized for structured output. Here's the pipeline:
- You describe your idea — "Create a product launch animation with a phone rotating in 3D, text fading in from below, and particles floating upward"
- AI generates a project spec — The model outputs a complete animation definition: canvas size, elements (text, shapes, 3D models), keyframes, easing curves, and timing
- KinoViz renders it — The editor parses the AI output and builds a fully editable project you can refine manually
What AI Can Generate
- Text animations: Titles, subtitles, captions with fade, slide, scale effects
- Shape animations: Rectangles, circles, custom paths with motion and color transitions
- Bubble effects: Speech bubbles, thought bubbles with entrance animations
- 3D camera movements: Orbit, zoom, pan on imported 3D models
- Auto music: Synchronized background tracks in 5 styles
Credit System
Each AI generation uses approximately 16K tokens and costs 1 credit. Credits are included in subscription plans:
| Plan | Monthly Credits | Extra Credits |
|---|---|---|
| Free | 5 | — |
| Creator | 50 | $0.10/credit |
| Pro | 200 | $0.08/credit |
| Studio | Unlimited | Included |
Tips for Better Results
- Be specific — "A 16:9 social media ad with bold white text on blue background" works better than "make something cool"
- Mention element types — Include words like "text", "shape", "3D model", "bubble" to guide the AI
- Describe timing — "Text appears at 0.5s, logo fades in at 1.2s" helps the AI sequence correctly
- Iterate — Generate a base, then fine-tune keyframes manually in the editor
What's Next
We're continuously improving the AI model with more training data and better structured output. Coming soon: multi-scene animations, voice-over generation, and style transfer from reference videos.
