What Is AI Video Generation?

AI video generation transforms static images into videos with realistic motion. The AI analyzes the scene and adds natural movement based on your prompt. A beach scene might show waves crashing and palm trees swaying. A product shot might show fabric texture moving or light reflecting.

This feature is part of the 4-phase video production pipeline:

  1. Scene Images - AI-generated visuals from your storyboard
  2. Audio Assets - Professional voiceover and original music
  3. Video Clips - AI animation from still images
  4. Final Export - Multi-aspect video composition

How to Enable AI Video

When creating videos in Video Production, you choose between two animation modes:

Ken Burns Mode

  • Pan and zoom on static images
  • Professional appearance without AI generation
  • Faster production (1-3 minutes)

AI Video Mode

  • Full motion animation from still images
  • Dynamic, eye-catching movement
  • Longer production (3-8 minutes)

To use AI Video Mode:

  1. Create a video script in Video Production
  2. Generate storyboard with scene images
  3. Approve scenes or regenerate as needed
  4. Select "AI Video Mode" in production settings
  5. Choose your preferred video quality tier
  6. Click "Produce Video"

The system automatically generates animated clips for each scene and composites them into your final video.

Video Generation Capabilities

ILLIXIS supports multiple AI video generation tiers to balance quality, speed, and budget:

Standard Tier

  • Aspect ratio: 16:9 (horizontal)
  • Duration: 6-20 seconds per clip
  • Resolution: 1080p (25fps)
  • Best for YouTube videos, website hero sections, horizontal social posts
  • Vertical and square formats are produced by center-cropping the horizontal output

Premium Tier

  • Aspect ratios: 16:9, 9:16, 1:1 (any format generated natively)
  • Duration: 3-15 seconds in 1-second increments
  • Resolution: 1080p
  • Native audio generation with lip-sync (optional)
  • Best for multi-platform campaigns requiring vertical, square, and horizontal formats from the same script
  • Highest motion quality

Mid Tier

  • Durations: 5 or 10 seconds
  • All aspect ratios supported
  • Native audio with lip-sync

Budget Tier

  • Durations: 6 or 10 seconds
  • All aspect ratios supported
  • Good motion quality at lower cost

Choosing the Right Tier

| Output Format | Recommended | Why |
|---------------|-------------|-----|
| YouTube only | Standard | Native 16:9, lowest cost for horizontal |
| TikTok/Reels only | Premium | Native 9:16, precise timing |
| Multi-platform | Premium | Generates all formats natively |
| Budget-conscious | Budget | Good quality, significantly cheaper |
| Lip-sync required | Premium or Mid | Native audio with speech sync |

How Duration Works

Each scene in your script gets its own video clip. Duration is determined by the voiceover text length:

Automatic Duration Calculation:

  • ILLIXIS estimates speaking time from text
  • Rounds to the nearest valid duration for the chosen tier
  • Standard tier: rounds to 6, 8, 10, 12, 14, 16, 18, or 20 seconds
  • Premium tier: uses exact duration between 3-15 seconds
  • Budget tier: rounds to 6 or 10 seconds

Example:

  • Scene text: "This product revolutionizes your morning routine with its innovative design."
  • Estimated speaking time: 4.2 seconds
  • Standard duration: 6 seconds (rounds up)
  • Premium duration: 5 seconds (rounds up to nearest second)

Longer durations ensure voiceover doesn't feel rushed. Shorter durations reduce cost but may require faster speech.

Motion Prompts

AI video models animate based on a motion prompt. ILLIXIS generates these automatically, but understanding them helps you choose the right animation mode.

Motion prompt strategy:

  • Describes desired movement, NOT caption text
  • Focuses on natural phenomena (wind, water, light)
  • Includes style guidance (subtle, dynamic, cinematic)
  • Never includes text content (AI would try to render it as visuals)

Automatic motion styles:

Subtle "Gentle natural movements. Soft ambient motion like gentle breeze, light shimmer, subtle sway. High quality, smooth motion."

Dynamic "Flowing movement throughout the scene. Wind, water ripples, fabric movement. Vibrant lighting with energy."

Cinematic "Smooth camera movement with professional lighting. Gentle motion blur, natural shadows, depth."

Most video scripts use subtle or dynamic automatically based on content type. Educational content gets subtle motion. Promotional content gets dynamic.

Output Formats

AI-generated clips feed into the multi-aspect export pipeline. You produce three formats from one generation:

  • 9:16 - Vertical for TikTok, Instagram Reels, YouTube Shorts
  • 1:1 - Square for Instagram Feed, Facebook
  • 16:9 - Horizontal for YouTube, website embeds

Standard tier generates 16:9 natively, then crops to other formats. Premium tier generates each aspect ratio independently for better composition.

Production Time

AI video generation is the slowest phase of production:

Phase timing:

  1. Scene images: 30-60 seconds
  2. Voiceover + music: 30-45 seconds
  3. Video clips: 3-8 minutes (longest phase)
  4. Final export: 1-2 minutes

A 4-scene video requires 4 separate clip generations. These run in parallel (typically 4 simultaneous). Expect 3-8 minutes total for the video clips phase.

When to Use AI Video

Use AI video when:

  • Creating social media content (motion grabs attention)
  • Product videos need dynamic showcase (fabric, liquid, texture)
  • Competing in feed-based platforms where static images get scrolled past
  • Budget allows for premium production
  • Timeline permits 5-10 minute production

Use Ken Burns when:

  • Publishing to YouTube where longer-form content is expected
  • Budget is tight
  • Need fast turnaround (under 3 minutes)
  • Content is educational and doesn't require dynamic visuals
  • Static images with professional pan/zoom meet your needs

Both modes produce high-quality output. The choice is budget vs. engagement impact.

Vertical Video (Standard Tier)

Standard tier only generates 16:9 video. ILLIXIS handles vertical output by:

  1. Generating horizontal 16:9 clip
  2. Center-cropping to 9:16 during export phase
  3. Positioning captions and branding in the safe zone

This works well for most content but can crop important elements. If your video has critical details at the edges, use the Premium tier which generates 9:16 natively.

Production Dashboard

Track AI video generation in Video Production:

Status indicators:

  • Generating Images: Phase 1
  • Generating Audio: Phase 2
  • Generating Video Clips: Phase 3 (AI video happening now)
  • Compositing: Phase 4
  • Complete: Ready for download

During Phase 3, you'll see:

  • Current scene being generated
  • Time remaining estimate

If generation fails on one scene, the system retries up to 3 times before flagging the scene as failed. You can regenerate failed scenes individually.

Troubleshooting

Video generation times out

  • Scenes longer than 15 seconds may time out
  • Solution: Shorten scene duration or split into two scenes

Video looks static (not much motion)

  • Images with no natural motion sources (plain backgrounds) animate less
  • Solution: Use images with elements that can move (fabric, hair, water, foliage)

Cropped vertical videos cut off important elements

  • Standard tier generates 16:9, then crops to 9:16
  • Center crop may remove edge content
  • Solution: Use Premium tier for native 9:16 generation, or adjust scene composition

Generation fails repeatedly

  • The system may be under heavy load
  • Solution: Try a different quality tier, or wait 5 minutes and retry

Video doesn't match voiceover duration

  • Scene duration rounded to the tier's valid values
  • May create gaps or overlaps
  • Solution: Edit voiceover text to match rounded duration, or use Premium tier for precise timing

Settings and Preferences

Video settings persist per tenant. Configure once in Settings > Video Production:

  • Default quality tier: Standard, Premium, Mid, or Budget
  • Animation mode: Ken Burns or AI Video
  • Aspect ratios: Which formats to produce by default
  • Motion style: Subtle, dynamic, or cinematic

Individual videos can override these defaults at production time.

Related Features

  • Video Production Dashboard - Full storyboard workflow including AI video generation
  • Ken Burns Animations - Static image alternative to AI video
  • Voice Library - Custom voice cloning for voiceovers
  • Multi-Aspect Export - Produce all three formats from one generation

FAQ

Q: Can I use my own images instead of AI-generated scene images? Not yet. Custom image upload for video scenes is on the roadmap.

Q: Why does AI video cost more than Ken Burns? AI video generation requires advanced ML models that are computationally intensive. Ken Burns uses simpler rendering with static images.

Q: Can I generate audio directly from the video model? Premium and Mid tiers support native audio generation with lip-sync. This is optional. The default workflow uses ILLIXIS voiceover generation for better control and consistency.

Q: How many videos can I produce per month? Check your plan tier in Settings > Usage. Video quotas vary by tier:

  • Starter: 4 videos/month
  • Professional: 12 videos/month
  • Enterprise: 40 videos/month

Q: What happens if one scene fails to generate? The pipeline retries up to 3 times. If still failing, the scene is marked as failed. You can regenerate that scene individually, switch to Ken Burns mode for that scene, or choose a different quality tier.

Q: Can I mix Ken Burns and AI video in the same video? Not currently. All scenes use the same animation mode. This is a potential future enhancement.

Q: What if my video is longer than 20 seconds per scene? ILLIXIS splits long scenes into multiple clips. Each clip maxes out at the tier's limit. Clips are seamlessly joined during composition.

Ready to lose the stack?

One platform. You approve. ILLIXIS executes. Marketing that just happens.

Join the waitlistNo spam, everUnsubscribe anytime
First 20 founding members: 50% off any plan for your first year.

Marketing, Unstacked.