Toon MCP

🚀 TOON MCP Server

The MCP (Model Context Protocol) server for TOON (Token-Oriented Object Notation) encoding, which can reduce LLM token usage by 50 - 70% when sending structured data.

🚀 Quick Start

The TOON MCP Server is a powerful tool for optimizing data encoding and reducing LLM token usage. Here's how you can get started.

✨ Features

• TOON Encoding: TOON is a compact data format optimized for LLM input. It uses a header - based format, which can significantly reduce the number of tokens compared to JSON.
• Multiple Tools: It provides several useful tools such as toon_encode, toon_decode, toon_analyze, and toon_optimize_prompt.
• Cost Savings: By reducing token usage, it can save costs when working with LLMs at scale.

📦 Installation

Quick Start (npx - no install needed)

Add the following configuration to your MCP settings:

Claude Desktop (~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "toon": {
      "command": "npx",
      "args": ["-y", "toon-mcp-server"]
    }
  }
}

Claude Code (~/.claude/settings.json):

{
  "mcpServers": {
    "toon": {
      "command": "npx",
      "args": ["-y", "toon-mcp-server"]
    }
  }
}

Global Install

npm install -g toon-mcp-server

Then add the following to your MCP settings:

{
  "mcpServers": {
    "toon": {
      "command": "toon-mcp"
    }
  }
}

As Claude Code Skill

# Download the skill
curl -o ~/.claude/skills/toon.md https://raw.githubusercontent.com/elminson/toon-mcp/main/skills/toon.md

Then use /toon in Claude Code.

💻 Usage Examples

In Claude Desktop/Code (with MCP)

Just ask Claude to use the tools:

• "Encode this JSON to TOON: [...]"
• "Analyze how much I'd save converting this data to TOON"
• "Optimize this prompt for token efficiency"

Programmatic (Node.js)

const { ToonEncoder } = require('toon-mcp-server/src/toon-encoder');

// Encode
const data = [
  { id: 1, name: 'Test', price: 99.99 },
  { id: 2, name: 'Test 2', price: 149.99 },
];
const toon = ToonEncoder.encode(data);

// Get stats
const json = JSON.stringify(data);
const stats = ToonEncoder.getStats(json, toon);
console.log(stats.savings.percent); // "64.5%"

// Decode
const decoded = ToonEncoder.decode(toon);

📚 Documentation

Available Tools

toon_encode

Convert data to TOON format. Supported formats: JSON, CSV, TSV, XML, HTML tables, YAML

Input: [{"name":"Alice","age":30},{"name":"Bob","age":25}]
Output: [name,age]
        Alice,30
        Bob,25

toon_decode

Convert TOON back to JSON.

toon_analyze

Analyze data and show potential token/cost savings.

toon_optimize_prompt

Find data sections in a prompt and convert them to TOON automatically.

🔧 Technical Details

The TOON format uses a header - based approach. Instead of repeating field names for every object like in JSON, it defines field names in a header row, which can significantly reduce the number of tokens when sending structured data to LLMs.

📈 Benchmarks

Tested with OpenAI GPT - 4o - mini:

💰 Cost Savings at Scale

⚙️ When to Use TOON

✅ Best for:

• Arrays of objects with same structure (tables, lists, records)
• API responses, database results
• Large datasets sent to LLMs
• Cost optimization at scale

⚠️ Less effective for:

• Deeply nested, non - uniform data
• Small payloads (<5 items)
• Data with many unique field structures

🤝 Contributing

Pull requests welcome! Please open an issue first to discuss changes.

📄 License

MIT