AI Marketing Videos

Create professional marketing videos via inference.sh CLI.

Quick Start

Requires inference.sh CLI (belt). Install instructions

belt login

# Generate a product promo video
belt app run google/veo-3-1-fast --input '{
  "prompt": "Sleek product reveal video, smartphone emerging from light particles, premium tech aesthetic, commercial quality"
}'

Video Ad Types

Type	Duration	Platform
Bumper Ad	6 seconds	YouTube
Short Ad	15 seconds	Instagram, Facebook
Standard Ad	30 seconds	YouTube, TV
Explainer	60-90 seconds	Website, YouTube
Product Demo	30-60 seconds	All platforms

Marketing Video Templates

Product Launch

# Dramatic product reveal
belt app run google/veo-3 --input '{
  "prompt": "Cinematic product launch video, premium tech device floating in space, dramatic lighting, particles and light effects, Apple-style reveal, commercial quality"
}'

Brand Story

# Emotional brand narrative
belt app run google/veo-3-1 --input '{
  "prompt": "Brand story video showing diverse people connecting through technology, warm color grading, lifestyle montage, emotional and inspiring, commercial"
}'

Feature Highlight

# Focus on specific feature
belt app run bytedance/seedance-1-5-pro --input '{
  "prompt": "Close-up product feature demonstration, hands interacting with device, clean background, informative, tech commercial style"
}'

Testimonial Style

# Talking head testimonial
belt app run google/veo-3-1-fast --input '{
  "prompt": "Customer testimonial style video, person speaking to camera, neutral office background, professional lighting, authentic feel"
}'

Before/After

# Transformation reveal
belt app run google/veo-3-1-fast --input '{
  "prompt": "Before and after transformation video, split screen transition, dramatic reveal, satisfying comparison, commercial style"
}'

Complete Ad Workflows

30-Second Product Ad

# 1. Opening hook (0-3s)
belt app run google/veo-3-1-fast --input '{
  "prompt": "Attention-grabbing opening, product silhouette in dramatic lighting, building anticipation"
}' > hook.json

# 2. Problem statement (3-8s)
belt app run google/veo-3-1-fast --input '{
  "prompt": "Frustrated person dealing with common problem, relatable everyday situation, documentary style"
}' > problem.json

# 3. Solution reveal (8-15s)
belt app run google/veo-3-1-fast --input '{
  "prompt": "Product reveal with features highlighted, clean demonstration, solving the problem shown before"
}' > solution.json

# 4. Benefits showcase (15-25s)
belt app run google/veo-3-1-fast --input '{
  "prompt": "Happy customer using product, lifestyle integration, multiple quick cuts showing benefits"
}' > benefits.json

# 5. Call to action (25-30s)
belt app run google/veo-3-1-fast --input '{
  "prompt": "Product hero shot with space for text overlay, professional lighting, commercial ending"
}' > cta.json

# 6. Generate voiceover
belt app run infsh/kokoro-tts --input '{
  "prompt": "Tired of [problem]? Introducing [Product]. [Key benefit 1]. [Key benefit 2]. [Key benefit 3]. Get yours today.",
  "voice": "af_nicole"
}' > voiceover.json

# 7. Merge all clips with voiceover
belt app run infsh/media-merger --input '{
  "videos": ["<hook>", "<problem>", "<solution>", "<benefits>", "<cta>"],
  "audio_url": "<voiceover>",
  "transition": "crossfade"
}'

Instagram/TikTok Ad (15s)

# Vertical format, fast-paced
belt app run google/veo-3-1-fast --input '{
  "prompt": "Fast-paced product showcase, vertical 9:16, quick cuts, trending style, hook in first 2 seconds, satisfying visually, Gen-Z aesthetic"
}'

# Add trendy music
belt app run infsh/media-merger --input '{
  "video_url": "<video>",
  "audio_url": "https://trending-music.mp3"
}'

Explainer Video

# 1. Write script
belt app run openrouter/claude-sonnet-45 --input '{
  "prompt": "Write a 60-second explainer video script for a SaaS product. Include: hook, problem, solution, 3 key features, social proof, CTA. Make it conversational."
}' > script.json

# 2. Generate visuals for each section
SECTIONS=("hook" "problem" "solution" "feature1" "feature2" "feature3" "social_proof" "cta")

for section in "${SECTIONS[@]}"; do
  belt app run google/veo-3-1-fast --input "{
    \"prompt\": \"Explainer video scene for $section, motion graphics style, clean modern aesthetic, SaaS product\"
  }" > "$section.json"
done

# 3. Generate professional voiceover
belt app run infsh/kokoro-tts --input '{
  "prompt": "<full-script>",
  "voice": "am_michael"
}' > voiceover.json

# 4. Assemble final video
belt app run infsh/media-merger --input '{
  "videos": ["<hook>", "<problem>", "<solution>", ...],
  "audio_url": "<voiceover>",
  "transition": "fade"
}'

Platform-Specific Formats

Facebook/Instagram Feed

belt app run google/veo-3-1-fast --input '{
  "prompt": "Square format product video 1:1, eye-catching visuals, works without sound, text-friendly, scroll-stopping"
}'

YouTube Pre-Roll

belt app run google/veo-3-1-fast --input '{
  "prompt": "YouTube ad style, skip button awareness (hook in 5 seconds), 16:9, professional commercial quality"
}'

LinkedIn

belt app run google/veo-3-1-fast --input '{
  "prompt": "Professional B2B product video, corporate style, clean and modern, business audience, subtle motion"
}'

TikTok/Reels

belt app run google/veo-3-1-fast --input '{
  "prompt": "TikTok native style ad, vertical 9:16, raw authentic feel, not overly polished, trendy, user-generated content aesthetic"
}'

Ad Creative Best Practices

Hook Formula (First 3 Seconds)

belt app run google/veo-3-1-fast --input '{
  "prompt": "Opening hook: [choose one]
  - Surprising visual transformation
  - Bold statement text animation
  - Relatable problem scenario
  - Curiosity gap visual
  - Satisfying action"
}'

Visual Hierarchy

1. Product hero - Clear, prominent
2. Benefits - Illustrated, not just stated
3. Social proof - Visible testimonials/numbers
4. CTA - Clear space for text overlay

Sound Design

# Add appropriate music
belt app run infsh/ai-music --input '{
  "prompt": "Upbeat commercial background music, modern, energetic, 30 seconds"
}' > music.json

belt app run infsh/media-merger --input '{
  "video_url": "<ad-video>",
  "audio_url": "<music>",
  "audio_volume": 0.5
}'

A/B Testing Variants

# Generate multiple creative variants
HOOKS=(
  "Problem-focused opening"
  "Product reveal opening"
  "Testimonial opening"
  "Statistic opening"
)

for hook in "${HOOKS[@]}"; do
  belt app run google/veo-3-1-fast --input "{
    \"prompt\": \"Marketing video with $hook, professional commercial quality\"
  }" > "variant_${hook// /_}.json"
done

Video Ad Checklist

• Hook in first 3 seconds
• Works without sound (captions/text)
• Clear product visibility
• Benefit-focused messaging
• Single clear CTA
• Correct aspect ratio for platform
• Brand consistency
• Mobile-optimized

Related Skills

# Video generation
npx skills add inference-sh/skills@ai-video-generation

# Image generation for thumbnails
npx skills add inference-sh/skills@ai-image-generation

# Text-to-speech for voiceover
npx skills add inference-sh/skills@text-to-speech

# Social media content
npx skills add inference-sh/skills@ai-social-media-content

# Full platform skill
npx skills add inference-sh/skills@infsh-cli

Browse all apps: belt app list