Claude Repurpose

Content Repurposing Engine for Claude Code -- Turn 1 piece of content into 10+ platform-optimized posts in seconds.

What It Does

Takes any content (YouTube video, blog post, podcast, local file) and generates platform-native outputs for:

Why Not Just Cross-Post?

Cross-posting the same text everywhere kills engagement. Each platform has different:

• Character limits (Twitter: 280 vs LinkedIn: 3,000)
• Algorithm preferences (Twitter rewards replies 27x, LinkedIn rewards saves 5x)
• Audience expectations (LinkedIn: professional, Reddit: peer-to-peer, TikTok: entertaining)
• Content formats (LinkedIn PDF carousels: 6.60% engagement, Instagram mixed-media: 2.33%)

This skill adapts your message for each platform, not just reformats it.

Install

# One-liner
curl -fsSL https://raw.githubusercontent.com/AgriciDaniel/claude-repurpose/main/install.sh | bash

# Or clone and install
git clone https://github.com/AgriciDaniel/claude-repurpose.git
cd claude-repurpose && bash install.sh

Optional: Install /banana for AI-generated quote cards and carousel cover images.

Usage

# Repurpose a blog post to all platforms
/repurpose https://example.com/blog/my-article

# Repurpose a YouTube video
/repurpose https://youtube.com/watch?v=abc123

# Only specific platforms
/repurpose https://example.com/post --platforms twitter,linkedin,instagram

# Quick mode (top 3 platforms only)
/repurpose https://example.com/post --brief

# With AI-generated images (/banana required)
/repurpose https://example.com/post --images

# Override brand voice
/repurpose https://example.com/post --voice witty

# Just analyze content (no outputs)
/repurpose analyze https://example.com/post

# Generate publishing calendar from outputs
/repurpose calendar

How It Works

Input (any content)
      |
      v
  Content Extraction        YouTube -> transcript
  (auto-detect type)        Blog URL -> article text
                            Audio -> Whisper transcription
                            Local file -> direct read
      |
      v
  Content Atomization       Extract 5-15 reusable atoms:
  (key insights, quotes,    stats, quotes, insights,
   statistics, arguments)   questions, contrarian takes
      |
      v
  6 Parallel Agents         Social (Twitter+Threads+LinkedIn+Facebook)
  (simultaneous)            Visual (Instagram+TikTok+Pinterest+Snapchat+Quotes)
                            Longform (Newsletter+Email+Reddit+Quora)
                            Community (YouTube+Skool+Discord)
                            Broadcast (WhatsApp+Telegram+Medium)
                            SEO (metadata across all)
      |
      v
  Output Directory          ./repurposed/<timestamp>/
  (30+ files organized      twitter/, linkedin/, instagram/,
   by platform)             facebook/, skool/, quotes/, etc.

Output Structure

./repurposed/2026-03-31-1430/
├── summary.md                 # Overview of all generated content
├── atoms.md                   # Extracted content atoms
├── twitter/
│   ├── thread.md              # 8-12 tweet thread
│   ├── standalone-tweet.md     # Standalone tweets
│   └── poll.md
├── linkedin/
│   ├── post.md
│   ├── carousel.md            # Slide-by-slide script
│   └── poll.md
├── instagram/
│   ├── carousel.md            # 7-10 slide script
│   ├── caption.md
│   └── reel-script.md
├── threads/
│   ├── thread.md              # 5-10 post thread
│   ├── standalone-posts.md    # 3-5 standalone posts
│   └── image-post.md          # Image post concept
├── tiktok/
│   ├── video-script.md        # 15-60s video script
│   ├── carousel.md            # Photo mode (2-10 slides)
│   └── stitch-duet.md         # Stitch/duet concept
├── pinterest/
│   ├── pins.md                # 3-5 pin descriptions
│   ├── idea-pin.md            # Idea pin script (5-10 slides)
│   └── boards.md              # Board suggestions
├── snapchat/
│   ├── story-script.md        # 3-5 frames (10s each)
│   ├── spotlight-script.md    # Up to 60s Spotlight
│   └── ar-concept.md          # AR lens concept
├── facebook/
│   ├── post.md
│   ├── poll.md
│   └── story.md
├── youtube-community/
│   ├── post.md
│   ├── image-concept.md
│   └── poll.md
├── skool/
│   ├── discussion.md
│   ├── challenge.md
│   └── poll.md
├── discord/
│   ├── announcement.md        # Server announcement
│   ├── thread-prompt.md       # Discussion thread starter
│   └── embed.md               # Rich embed message
├── newsletter/
│   ├── newsletter.md          # Newsletter excerpt
│   ├── subject-lines.md
│   └── email-sequence.md
├── reddit/
│   ├── post.md
│   └── subreddits.md
├── quora/
│   ├── answer.md              # 300-1000 word answer
│   ├── space-post.md          # Quora Space post
│   └── questions.md           # Question suggestions
├── medium/
│   ├── article.md             # 1500-3000 word article
│   ├── tags-publications.md   # Tags + publication suggestions
│   └── crosspost-note.md      # Canonical URL handling
├── whatsapp/
│   ├── update.md              # Channel update (100-300 chars)
│   ├── poll.md                # Channel poll
│   └── teaser.md              # Content teaser with link
├── telegram/
│   ├── post.md                # Channel post (500-1000 chars)
│   ├── deep-dive.md           # Formatted deep dive
│   └── poll.md                # Poll with optional quiz mode
├── quotes/
│   ├── quotes.md
│   └── banana-prompts.md
├── seo-metadata.md
├── calendar.md
└── images/                    # AI-generated (with --images)

Platform Intelligence (2026 Data)

Built-in knowledge of each platform's current algorithm and best practices:

Image Generation (/banana)

With Claude Banana installed, the --images flag generates:

• Quote cards (5): Minimalist designs with content's best quotes
• Carousel covers: Bold title slides for Instagram and LinkedIn
• Hero images: Topic-relevant visuals for social posts

Without /banana, prompts are saved to banana-prompts.md for manual use.

Requirements

• Claude Code CLI
• Python 3.10+
• Optional: yt-dlp (YouTube transcript extraction)
• Optional: Whisper (audio transcription)
• Optional: /banana (AI image generation)

Architecture

• 20 sub-skills covering every platform
• 6 parallel subagents for maximum speed
• 8 reference files with platform specs, hook formulas, voice rules, and benchmarks
• 3 Python scripts for content extraction (article, transcript, audio)

Ecosystem

Part of the AgriciDaniel Claude Code skill family:

License

MIT

Author

Built by Agrici Daniel - AI Workflow Architect.

• Blog - Deep dives on AI marketing automation
• AI Marketing Hub - Free community, 2,800+ members
• YouTube - Tutorials and demos
• All open-source tools