🚀 AITokenSave-MCP

Smart copy-paste operations for AI workflows - Save 60-90% tokens by avoiding full file regeneration

🎯 Why AITokenSave-MCP?

Ever needed to modify just a few lines of code but AI regenerated the entire file? The traditional approach wastes massive amounts of tokens:

• Manual copy-paste - Loses formatting and requires complex prompts
• Full file regeneration - AI rewrites 200 lines to change 3 lines
• No precise targeting - "Add a function after line 42" becomes guesswork
• Indentation nightmares - Manual formatting breaks code structure

AITokenSave-MCP solves this:

• ✅ Extract specific line ranges with surgical precision
• ✅ Insert content at exact positions with auto-indentation
• ✅ Replace code blocks without touching surrounding content
• ✅ Save 60-90% tokens in typical code modification tasks

⚡ How It Works

1. Tell AI What You Need

"Add this Express route after line 15"

2. AI Uses Precise Tools

{
  "tool": "insert-at-line",
  "line": 15,
  "content": "app.get('/api/users', ...)"
}

3. Perfect Result

• ✅ Exact positioning
• ✅ Auto-indented code
• ✅ Zero regeneration
• ✅ 90% fewer tokens

📊 Token Savings

Task	Before (Full Regeneration)	After (Precise Tools)	Savings
Add function	2000+ tokens	200 tokens	90%
Replace config	1500+ tokens	150 tokens	90%
Extract snippet	1200+ tokens	100 tokens	92%

🚀 Installation

Option 1: NPX (Instant Use)

npx aitoken-save-mcp

Option 2: Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "aitoken-save": {
      "command": "npx",
      "args": ["aitoken-save-mcp"],
      "transport": "stdio"
    }
  }
}

Option 3: Global Install

npm install -g aitoken-save-mcp
aitoken-save-mcp

🛠️ Available Tools

extract-lines - Copy specific ranges

{
  "text": "source code",
  "start_line": 10,
  "end_line": 20
}

insert-at-line - Insert with auto-indentation

{
  "text": "original code",
  "content": "new code",
  "line": 25,
  "replace_lines": 3
}

copy-lines - Enhanced extraction

{
  "text": "source code", 
  "start_line": 5,
  "end_line": 15,
  "trim": true
}

💡 Usage Examples

Express Route: insert-at-line with exact positioning → 90% token savings Config Update: replace-lines targeting specific block → 87% token savings
Extract Function: extract-lines with perfect preservation → 85% token savings

Quick Patterns

# Copy-paste workflow
extract-lines: start_line=45, end_line=60
insert-at-line: line=25, content=extracted_code

# Smart positioning
line: 0     # Beginning    line: -1    # End    line: 42   # After line 42

🏗️ Design & Development

Minimal by Design: Only 3 tools, zero config, auto-indentation Built for Efficiency: ~50KB bundle, <5s install, memory efficient

# Development setup
git clone <repository-url> && cd aitoken-save-mcp
npm install && npm run dev

❓ FAQ

Auto-indentation: Detects surrounding code style automatically Line numbering: 1-based indexing (line 1 = first line)
Language support: Works with any programming language MCP compatibility: Claude Desktop, Cursor, VS Code, etc.

🤝 Contributing

• Keep tools minimal and focused
• Auto-indentation must always work
• No breaking changes to existing tools
• Comprehensive examples for new features

📄 License

MIT License - Free for commercial and personal use

---

⭐ Stop wasting tokens on full file regeneration - Start using precision tools!

Built for efficiency. Designed for developers. Optimized for AI workflows. 🤖✨