Skill: Video Hook Generation

Purpose

Create 10-30 second video hooks for social media (TikTok, Instagram Reels, YouTube Shorts). These are the primary discovery mechanic for AI music — short, visually striking clips that stop the scroll.

Trigger

When the user needs social media video content for a song, or when /generate-assets requests video hooks.

Hook Types

Type 1: Animated Cover (10s)

The simplest and fastest to produce. Take the album cover and add subtle motion:

• •Floating particles matching the song's energy
• •Gentle camera drift (slow zoom in or pan)
• •Pulsing light effects synced to BPM
• •Waveform visualization overlay

Tools: Nano Banana (image gen) → CapCut/DaVinci (animation) or CSS animation for web Best for: Audio previews, Spotify Canvas, Instagram story

Type 2: Lyric Hook (15-20s)

The chorus or hook line animated with kinetic typography:

• •Extract the catchiest 2-4 lines
• •Animate text revealing word by word, synced to vocals
• •Background: album cover or abstract visual matching mood
• •Add audio waveform at bottom

Tools: CapCut auto-captions → style with brand fonts Best for: TikTok, Instagram Reels

Type 3: Visual Story (20-30s)

A mini-narrative with 3-4 scenes:

• •Scene 1 (0-8s): Establishing shot — mood setter, no vocals
• •Scene 2 (8-16s): Build — introduce the hook visually
• •Scene 3 (16-24s): Climax — the chorus hits, visual intensity peaks
• •Scene 4 (24-30s): Resolution — artist name, "Full song on Spotify/Suno"

Tools: AI image gen for each scene → video editor to stitch + add audio Best for: YouTube Shorts, TikTok, Instagram Reels

Type 4: Behind-the-Scenes (15-30s)

Show the creative process:

• •Screen recording of Suno generating the song
• •Split screen: prompt on left, waveform on right
• •Text overlay: "I made this with AI in 2 minutes"
• •Result plays with full audio

Tools: Screen recording → CapCut editing Best for: Creator/tech audience, TikTok

Scene Generation Guide

Image Prompts for Video Scenes

For each scene in a Visual Story hook, generate an image with:

Scene prompt structure:
[Mood descriptor], [specific scene description],
[character/subject], [lighting], [camera angle],
cinematic still frame, 16:9 aspect ratio,
movie screenshot quality, [genre-specific style]

Camera Angles by Energy Level

Scene-to-Song Timing

BPM → Beats per second → Cut timing

60 BPM  = 1 beat/sec → cuts every 4-8 beats (4-8 sec)
120 BPM = 2 beats/sec → cuts every 8-16 beats (4-8 sec)
140 BPM = 2.3 beats/sec → cuts every 8-12 beats (3-5 sec)
170 BPM = 2.8 beats/sec → cuts every 8-16 beats (3-6 sec)

Platform Specifications

TikTok / Instagram Reels

• •Aspect: 9:16 (1080x1920)
• •Duration: 15s or 30s (15s preferred for hooks)
• •First 3 seconds: Must grab attention — start with the hook, not the intro
• •Text safe zone: Keep important text in center 80%
• •Audio: Must start with the catchiest part of the song

YouTube Shorts

• •Aspect: 9:16 (1080x1920)
• •Duration: Up to 60s (but 30s is optimal)
• •Thumbnail: Auto-generated from frame, so make every frame good
• •Audio: Can use full song clip

Spotify Canvas

• •Aspect: 9:16
• •Duration: 3-8 seconds, seamless loop
• •No text allowed — visual only
• •Format: MP4, under 10MB

Instagram Story

• •Aspect: 9:16 (1080x1920)
• •Duration: Up to 15s per story
• •Include: Swipe-up/link sticker to full song

Workflow

Step 1: Analyze Song

Input: Song metadata (genre, BPM, mood, lyrics excerpt, persona)
Output: Hook type recommendation + scene breakdown

Step 2: Generate Scene Images

For Visual Story hooks, generate 3-4 images:

For each scene:
1. Compose image prompt (see Scene Generation Guide)
2. Generate via Nano Banana or FLUX
3. Ensure visual consistency across scenes (same color palette, style)

Step 3: Create Assembly Guide

Output a detailed cut sheet:

HOOK: "Trust in Yourself" — 20s Visual Story

Scene 1 (0:00-0:06): Wide shot — empty arena at dawn, spotlights warming up
  Audio: Instrumental intro, building tension
  Motion: Slow zoom into center stage

Scene 2 (0:06-0:12): Medium shot — figure walking toward spotlight
  Audio: First verse begins
  Motion: Tracking shot, camera follows from behind

Scene 3 (0:12-0:18): Close-up — face illuminated, singing the chorus
  Audio: "TRUST IN YOURSELF" chorus hits
  Motion: Quick zoom to eyes, light burst

Scene 4 (0:18-0:20): Wide — full arena lit up, crowd silhouettes
  Audio: Chorus continues
  Text overlay: "Trust in Yourself — FrankX | Full song on Suno"

Step 4: Export Instructions

Provide platform-specific export settings for CapCut/DaVinci:

• •Resolution, FPS, codec
• •Where to add the audio (timestamp + fade)
• •Text overlay positioning

CapCut Project Generation

For users with CapCut, generate a project structure:

{
  "project": "Trust in Yourself — Hook",
  "resolution": "1080x1920",
  "fps": 30,
  "duration": "20s",
  "tracks": [
    {
      "type": "audio",
      "file": "trust-in-yourself.mp3",
      "startTime": 0,
      "duration": 20,
      "fadeIn": 0.5,
      "fadeOut": 1.0
    },
    {
      "type": "video",
      "clips": [
        { "file": "scene-1.png", "start": 0, "end": 6, "animation": "slow-zoom" },
        { "file": "scene-2.png", "start": 6, "end": 12, "animation": "ken-burns-right" },
        { "file": "scene-3.png", "start": 12, "end": 18, "animation": "zoom-burst" },
        { "file": "scene-4.png", "start": 18, "end": 20, "animation": "fade-out" }
      ]
    },
    {
      "type": "text",
      "content": "Trust in Yourself — FrankX",
      "startTime": 18,
      "duration": 2,
      "position": "bottom-center",
      "font": "Inter Bold",
      "color": "#ffffff"
    }
  ]
}

Quality Rules

• •First 3 seconds MUST be visually striking — never start with black/silence
• •Audio MUST start with the hook, not the intro (unless instrumental)
• •Text must be readable on mobile (min 48px on 1080 wide)
• •Color palette must match album cover
• •Each scene must be a different composition (no repeated shots)
• •Total file size under 50MB for upload