How AI Generation Works

Understand what happens behind the scenes when iArt generates your video.

Written By Yunfei

Last updated 1 day ago

The Generation Process

When you submit a prompt, here's what happens:

  1. Prompt analysis — The AI agent reads your description and understands what you want to create

  2. Code generation — The agent writes animation code using Remotion and GSAP, translating your text into motion graphics

  3. Compilation and preview — The generated code is compiled and rendered in your browser for instant preview

This entire process typically takes 15–30 seconds, depending on the complexity of your request.

What the AI Can Create

The AI agent is capable of generating a wide range of animations:

  • Text animations (typing effects, fade-ins, slides, bounces)

  • Shape and graphic animations

  • Color transitions and gradients

  • Layout compositions with multiple elements

  • Sequential and staggered animations

  • Smooth transitions between scenes

Generation Time

ComplexityTypical Time

Simple (title + subtitle)

10–15 seconds

Medium (3–5 animated elements)

15–25 seconds

Complex (multiple sections, detailed animations)

25–40 seconds

Why Results May Vary

Each generation is a fresh AI interpretation of your prompt. This means:

  • Two identical prompts may produce slightly different results — this is normal

  • More specific prompts produce more consistent results

  • Vague prompts give the AI more creative freedom, leading to more variation

Re-generating

If the result doesn't match your expectations, you have two options:

  • Refine through chat — Describe what you'd like to change. This is usually the faster approach since the AI builds on the existing work

  • Re-generate — Submit the same or modified prompt to get a completely fresh result

Next Steps