AI Video Clipper

Transform long-form YouTube videos or local video files into viral-ready short clips for Instagram Reels and TikTok using AI-powered analysis.

---

Quick Start (Choose One)

Option 1: Use with an AI Assistant (Easiest)

Give the skill file to your AI assistant (Claude Code, Codex, etc.):

"Install this skill: universal-video-clipper.skill"

Then simply ask:

“Create viral clips from this YouTube video: https://www.youtube.com/watch?v=VIDEO_ID”

Option 2: Use with Claude Code / Codex

Tell your AI assistant to install this project:

"Install this project into your skill catalog"

The AI will automatically use the skill when you ask for video clips.

Option 3: Run Directly (No AI)

python ai_clip_generator.py "https://www.youtube.com/watch?v=VIDEO_ID"
python ai_clip_generator.py "/path/to/video.mp4"

---

How to Use This Tool

Once installed (or running directly), use natural language:

For YouTube Videos

“Create viral clips from this YouTube video: https://www.youtube.com/watch?v=VIDEO_ID”

“Turn this YouTube video into Instagram Reels: [URL]”

“Create viral clips from this YouTube video with animated captions: [URL]”

“Extract the best highlights from this tutorial for TikTok”

For Local/Uploaded Videos

“Create short clips from this video file: /path/to/video.mp4”

“Make Reels from my uploaded video”

“Make Reels from my uploaded video with scaling caption style”

With Additional Options

“Create 5 viral clips from this video with animated captions”

“Extract highlights from this YouTube video, skip downloading since I already have it”

---

What This Tool Does

1. Analyzes your video - Downloads (YouTube) or ingests (local files) the source video
2. Transcribes audio - Uses Whisper to create timestamped text transcripts
3. Finds viral moments - AI analyzes the content for engaging clips (hooks, value bombs, demos)
4. Generates clips - Creates 15-60 second clips optimized for social media
5. Formats for platforms - Outputs both standard and 9:16 vertical versions

---

What You Get

For each video processed, you’ll receive:

• Standard clips - Original aspect ratio versions
• Instagram versions - 9:16 vertical format (1080x1920) with padding
• Captioned variants - Optional animated word-by-word captions (CapCut-style)
• Summary report - Suggested captions, hashtags, virality scores

Sample Output

downloads/{video_id}/
├── original.mp4                    # Source video
├── original.json                   # Transcript with timestamps
├── metadata.json                   # Video info
├── analysis_request.md             # AI analysis prompt
├── clip_recommendations.json       # AI-generated clip suggestions
├── SUMMARY.md                      # Final report with captions & hashtags
└── clips/
    ├── clip_001_hook.mp4           # Standard clip
    ├── clip_001_hook_instagram.mp4 # 9:16 vertical
    ├── clip_001_hook_captioned.mp4 # With animated captions
    └── ...

---

Quick Example

YouTube Input

Input: https://www.youtube.com/watch?v=VIDEO_ID

Output: 3-5 ready-to-post clips with:

• Virality scores (1-10)
• Suggested captions with emojis
• Hashtag recommendations
• Standard + Instagram versions

Local File Input

Input: /Users/you/Videos/my_recording.mp4

Output: Same as above - works with any MP4, MOV, or video file on your computer

---

Prerequisites

Before first use, ensure these are installed:

Verify installation:

ffmpeg -version
yt-dlp --version
whisper --help
cd remotion-captions && npx remotion --help

---

Technical Reference

CLI Command (Alternative to Natural Language)

If you prefer direct script execution:

# YouTube video
python ai_clip_generator.py "https://www.youtube.com/watch?v=VIDEO_ID"

# Local video file (MP4, MOV, etc.)
python ai_clip_generator.py "/path/to/video.mp4"
python ai_clip_generator.py "~/Movies/my_recording.mov"

# With options
python ai_clip_generator.py "URL" --add-captions
python ai_clip_generator.py "/path/to/video.mp4" --add-captions
python ai_clip_generator.py "URL" --skip-download

Caption Styles

Skip Flags

# Re-run with existing files
python ai_clip_generator.py "URL" --skip-download
python ai_clip_generator.py "URL" --skip-download --skip-transcription

---

Workflow Details

Step 1: Video Ingest

• YouTube URLs: Downloads via yt-dlp (tries Chrome cookies → Firefox → no cookies)
• Local files: Copies to working directory (supports MP4, MOV, MKV, and most video formats)

Step 2: Transcription

• Uses OpenAI Whisper for speech-to-text
• Generates timestamps for each segment
• Falls back from “base” to “tiny” model if needed

Step 3: AI Analysis

The script pauses and waits for AI analysis:

1. Review analysis_request.md (contains transcript + instructions)
2. Run AI on this prompt to generate clip recommendations
3. Save results as clip_recommendations.json
4. Press Enter to continue

Step 4: Clip Generation

• Validates clip timestamps
• Uses FFmpeg to extract segments
• Creates dual formats: standard + Instagram 9:16

Step 5: Summary Report

• Generates SUMMARY.md with captions, hashtags, virality scores

---

JSON Recommendation Format

When AI analyzes the transcript, it should return JSON like:

{
  "clips": [
    {
      "clip_number": 1,
      "start_time": 5.2,
      "end_time": 32.8,
      "duration": 27.6,
      "title": "Hook: AI Creates Apple Shortcuts",
      "description": "Opens with bold claim and immediate demonstration",
      "virality_score": 9,
      "virality_factors": ["strong_hook", "visual_demo", "trending_topic"],
      "suggested_caption": "🤯 I made AI create Apple Shortcuts!",
      "content_type": "hook"
    }
  ],
  "video_summary": "Brief summary of video content",
  "overall_theme": "AI Automation",
  "hashtag_suggestions": ["#AI", "#automation", "#tech"]
}

---

Troubleshooting

---

Understanding the Output

Virality Score (1-10)

• 8-10: Highly recommended, strong viral potential
• 6-7: Good, worth posting
• Below 6: Consider adjusting timestamps or skipping

Content Types

• hook - Attention-grabbing opening (0-3 seconds)
• value_bomb - Actionable tips, “aha!” moments
• demo - Technical demonstrations
• story - Complete narrative with payoff

---

Project Overview

This tool is designed for two types of users:

1. AI Agents - Use natural language prompts shown in the “How to Use This Tool” section above
2. Direct users - Run CLI commands shown in “Technical Reference” section

Features

• Smart clip detection based on viral content patterns
• Automated workflow from URL to final clips
• Social media optimization (15-60 second duration, 9:16 format)
• Comprehensive reporting with captions and hashtags

Phase 2 Features (Implemented)

• Animated caption generation using Remotion
• Word-by-word sync with Whisper timestamps
• Three customizable caption styles

---

AI Assistant Integration (Optional)

If you use an AI assistant (like Claude Code), you can import the packaged skill:

universal-video-clipper.skill - A packaged skill file (ZIP archive) containing:

• SKILL.md with instructions for the AI
• Reference documentation
• Helper scripts
• scripts/remotion-captions/ for animated caption rendering

To use it, import universal-video-clipper.skill into your AI assistant’s skill manager.

---

License

Open source - feel free to use, modify, and distribute.

---

Ready? Try: “Create viral clips from this YouTube video: https://www.youtube.com/watch?v=QE_Nt5dMLHI”